Anda di halaman 1dari 35

Personal Relationships, 5 (1998),357-391. Printed in the United States of America.

Copyright 0 1998 ISSPR. 1350-4126/98$9.50

The Investment Model Scale: Measuring


commitment level, satisfaction level,
quality of alternatives, and investment size

CARYL E. RUSBULT,O JOHN M. MARTZ? AND


CHRISTOPHER R. AGNEWb
aUniversity of North Carolina at Chapel Hill; and bPurdue University

Abstract
Three studies evaluated the reliability and validity of the Investment Model Scale, an instrument designed to
measure four constructs, including commitment level and three bases of dependence-satisfaction level,
quality of alternatives, and investment size. In all three studies, reliability analyses revealed good internal
consistency among items designed to measure each construct. Also, principal components analyses performed
on scale items revealed evidence of four factors, with items designed to measure each construct loading on
independent factors. Studies 2 and 3 examined associations of model variables with instruments measuring
diverse qualities of relationships and assorted personal dispositions. As anticipated, Investment Model
variables were moderately associated with other measures reflecting superior couple functioning (e.g., dyadic
adjustment, trust level, inclusion of other in the self), and were essentially unrelated to measures assessing
personal dispositions (e.g., need for cognition, self-esteem). In addition, Study 3 demonstrated that earlier
measures of Investment Model variables predicted later levels of dyadic adjustment and later relationship
status (persisted vs ended). It is hoped that the existence of a reliable and valid Investment Model Scale will
promote further research regarding commitment and interdependence in ongoing close relationships.

This research was supported in part by a grant to the first Ashley Hermon, Molly Horton, Susan Hunter, Janet
author from the NSF (No. BNS-9023817); preparation Kent, Elizabeth Leatherland, Janet Macatangay, Jen-
of this manuscript was supported in part by a Reynolds nifer Oldham, Jill Ruppe, Stacey Sigmon, Crystal
Research and Study Leave from the University of North Smith, LauraFrancis Thorp, Leslie Turner, Sarah Whit-
Carolina at Chapel Hill and by a Visitors Grant from ton, and Mary Zawasky for their help in conducting
the Netherlands Organization of ScientificResearch. Study 2; and we thank Nathan Bitner, Kim Fleischman,
We are grateful to Michael Wexler and Ximena Ar- Tim Fowler, Wendy Greenhouse, Jenny Hamer, Vickie
riaga for their generous assistancein datamanagement. Hannah, Linda Johnson, Marissa Keller-Gusman, Jun
We also thank Steffaney Batson, Kimberly Bolton, Sue Kim, Kenny Levine, Sophia Min, Theresa Ploe, Debbie
Busby, Julie Carter, Allison Cashwell, Michael Dubs, Rhatigan, and StefanTyson for their help in conducting
Yahonda Glover, Tara Green, Elizabeth Henderson, the Study 3 telephone interviews.
Rebecca Holt, Angela Jones, Sarah Jordan, Cheryl Christopher Agnew is in the Department of Psy-
Kaufman, David Kessel, Channing King, Jennifer Kur- chological Sciences, Purdue University, West La-
fees, Mary Ann Lee, Samantha McCafferty, April fayette, IN 47907-1364.
McLawhorn, Linda Parker, Dawn Pilger, Frank Pyne, Correspondence regarding the Investment Model
Heather Quel, John Rights, Elizabeth Ross, Diane Scale should be addressed to Caryl Rusbult, Depart-
Runk, Lisa Seago, Andrea Sears, Rebecca Seymour, ment of Psychology, University of North Carolina at
Traci Simpson, Christopher Smith, Shannon South, Chapel Hill, Chapel Hill, NC 27599-3270. E-mail: rus-
Melanie Stevenson, Cynthia Wellons, Kristin Whitley, bult@unc.edu
and Charles Winstead for their help in conducting This research is dedicated to John Martz (1964-
Study 1;we thank Kristie Barreau, JohnBeckwith, Julie 1995). John was an excellent scientist and an excep-
Bullock, khan Canty, Doug Davis, Jody Davis, Eliza- tional human being. His enthusiasm for science and for
beth Farmer, Lisa Fox, David Grady, Whitney Half- lifes other pleasures infected and inspired his friends,
penny, Buddy Hams, Lindon Hayes, Jennifer Helms, colleagues,and students. We miss him very much.

357
358 C.E. Rusbult, J.M. Martz, and C.R. Agnew

Over the past two decades social scientists ries of commitment processes have been
have exerted considerable effort toward advanced (Brickman, Dunkel-Schetter, &
understanding why some relationships per- Abbey, 1987; Johnson, 1991; Kelley, 1983;
sist over time whereas others wither and Levinger, 1979; Rusbult, 1980a). These
die. Many researchers have assumed that theories share the assumption that commit-
the best route to understanding persistence ment is a key issue in understanding why
is to explore the determinants and conse- some relationships persist and others do
quences of positive affect-attraction, satis- not. Among these theories, Rusbults
faction, or love. For example, the goal of (1980a) Investment Model has been shown
many studies is to explain the causes of at- to be especially powerful in predicting com-
traction or love, and measures of satisfac- mitment and persistence across many types
tion frequently are employed as indices of of romantic relationship (e.g., marital rela-
couple health (for reviews of the literature, tionships, lesbian and gay relationships) as
see Berscheid, 1994; Berscheid & Reis, well as in friendships and in organizational
1998). The implicit or explicit assumption is settings (for reviews of the literature, see
that if partners love each other and feel Rusbult, 1987; Rusbult & Buunk, 1993;
happy with their relationship, they will be Rusbult, Drigotas, & Verette, 1994).
more likely to persist in their relationship. Unfortunately, empirical research re-
In some respects this point of view garding the Investment Model has pro-
makes good sense: All things considered, it ceeded in a somewhat haphazard manner, in
is easier to persist when a relationship feels that no published instrument exists for
good than when it feels bad. But at the measuring commitment and its antecedents.
same time, it may be somewhat simplistic to This article presents the results of three
assume that happiness tells the whole story studies demonstrating the reliability and va-
in explaining persistence. Stripped to its es- lidity of the Investment Model Scale, an
sence, such a model of human persistence instrument designed to measure four key
suggests that as long as it feels good, Ill predictors of persistence, including commit-
stick with it. Importantly, this conventional ment level and three bases of depend-
focus on the study of affective reactions ence-satisfaction level, quality of alterna-
fails to answer three key questions: First, tives,and investment size.Tojustify the need
for such a scale, we begin by describing the
why do some relationships persist despite
Investment Model and reviewing existing
dissatisfaction; for example, why do un-
research regarding commitment processes.
happy partners sometimes remain together
due to inertia, or for the sake of the chil-
dren? Second, why do some satisfying re-
lationships end; for example, why do indi- Determinants of Commitment:
viduals sometimes abandon relatively Satisfaction, Alternatives, and Investments
happy relationships to pursue desirable al- The Investment Model emerged out of In-
ternatives? And third, how can we account terdependence Theory and employs inter-
for persistence in the face of ordinary fluc- dependence constructs to analyze the ten-
tuations in relationships; given that satisfac- dency to persist in a relationship (Kelley,
tion ebbs and flows even in the most grati- 1979; Kelley & Thibaut, 1978; Thibaut &
fying involvements, and given that tempting Kelley, 1959). Interdependence Theory is a
alternatives threaten even the most smitten unique orientation in that its explanatory
partners, why do some relationships survive power rests on an analysis of the interde-
such fluctuations whereas others do not? pendence structure characterizing a given
Recognizing that whether a relationship relationship, not on the personal disposi-
is satisfying and whether it persists to some tions of the involved persons. Depen-
degree may be separate issues, several theo- dence is a central feature of interdepen-
Investment model scale 359

dence structure, particularly insofar as we respects (Rusbult, 1980a, 1983). First, the
seek to understand persistence in a rela- Investment Model suggests that satisfaction
tionship. Level of dependence refers to the level and alternative quality do not fully
extent to which an individual needs a explain dependence. If dependence was
given relationship, or relies uniquely on the based solely on the satisfactions derived
relationship for attaining desired outcomes. from the current relationship in compari-
How do individuals become dependent son to those anticipated elsewhere, few re-
on their relationships? Interdependence lationships would endure-a relationship
Theory identifies two main processes would falter on the occasion of poor out-
through which dependence grows. First, and comes or the appearance of an attractive
consistent with the fields traditional em- alternative. In reality, some relationships
phasis on positive affect, individuals be- survive even when an attractive alternative
come increasingly dependent to the extent is available, and even when a relationship is
that they experience high satisfaction in a not very gratifying. How can we explain
relationship. Satisfaction level refers to the persistence in the face of tempting alterna-
positive versus negative affect experienced tives and fluctuating satisfaction?
in a relationship. Satisfaction is influenced The Investment Model asserts that de-
by the extent to which a partner fulfills the pendence is also influenced by a third fac-
individuals most important needs. For ex- tor-investment size. Investment size refers
ample, Bill is likely to feel satisfied to the to the magnitude and importance of the
degree that Mary gratifies his intellectual, resources that are attached to a relation-
companionate, and sexual needs. ship-resources that would decline in value
However, satisfaction is not the sole ba- or be lost if the relationship were to end (cf.
sis for dependence. According to Inter- Becker, 1960; Rubin & Brockner, 1975;
dependence Theory, dependence is also Staw, 1976;Teger, 1980;Tropper, 1972).As a
influenced by the quality of available alter- relationship develops, partners invest many
natives. Quality of alternatives refers to the resources directly into their relationship in
perceived desirability of the best available the hope that doing so will improve it. For
alternative to a relationship. Quality of al- exampie, Bill may disclose his private
ternatives is based on the extent to which thoughts and feelings to Mary, and may put
the individuals most important needs considerable time and effort into their rela-
could effectively be fulfilled outside of tionship. Moreover, some investments are
the current relationship-in a specific alter- indirect, and come into existence when
native involvement, by the broader field of originally extraneous resources such as mu-
eligibles, by friends and family members, or tuat friends, personal identity, children, or
on ones own. For example, to the degree shared material possessions become at-
that Bills needs for intimacy and compan- tached to a relationship. Invested resources
ionship could not be gratified elsewhere, presumably enhance commitment because
quality of alternatives is poorer and his de- the act of investment increases the costs of
pendence on Mary is greater. ending a relationship, serving as a powerful
Thus, Interdependence Theory suggests psychological inducement to persist.
that dependence on a relationship is The Investment Model further extends
greater to the extent that an individual Interdependence Theory by suggesting that
wants to persist with a given partner (i.e., feelings of commitment emerge as a conse-
satisfaction level is high), and to the extent quence of increasing dependence. Commit-
that an individual has no choice but to per- ment level is defined as intent to persist in a
sist with that partner (i.e., alternatives are relationship, including long-term orienta-
poor). The Investment Model extends In- tion toward the involvement as well as feel-
terdependence Theory propositions in two ings of psychological attachment (e.g., a
360 C.E. Rusbult, J.M. Martz, and C.R.Agnew

Satisfaction

Probability of

Investment
Size

Figure 1. The investment model of commitment processes.

sense of we-ness; Agnew, Van Lange, sions to persist-that is, commitment medi-
Rusbult, & Langston, 1998). How does ates the effects on persistence of the three
commitment differ from dependence? De- bases of dependence (see Figure 1).
pendence is a fundamental quality of rela-
tionships-a relationship state describing
the additive effects of wanting to persist Consequences of Commitment:
(feeling satisfied), needing to persist (hav- Persistence and Relationship
ing high investments), and having no choice Maintenance Mechanisms
but to persist (possessing poor alternatives; The empirical literature provides consistent
see Figure 1). As individuals become in- support for Investment Model claims, dem-
creasingly dependent they tend to develop onstrating that (a) commitment is posi-
strong commitment. Commitment can be tively associated with satisfaction level and
construed as a sense of allegiance that is investment size, and is negatively associ-
established with regard to the source of ated with quality of alternatives; (b) each of
ones dependence. For example, because these variables contributes unique variance
Bill is dependent on his relationship with to predicting commitment; (c) compared to
Mary, Bill develops an inclination to persist less committed individuals, highly commit-
with Mary, he comes to think of himself as ted individuals are substantially more likely
part of BillandMary, and he considers the to persist in their relationships; and (d)
broader implications of his actions-impli- commitment is the most direct and power-
cations extending beyond his immediate ful predictor of persistence, partially or
self-interest, including effects on the rela- wholly mediating the effects of satisfaction,
tionship next week and next month and alternatives, and investments on decisions
next year. As such, the psychological expe- to remain in versus end a relationship. Such
rience of commitment reflects more than findings have been observed in several cul-
the bases of dependence out of which it tures (e.g., the United States, the Nether-
arises. Commitment is the psychological lands, and Taiwan), in research employing
construct that directly influences everyday diverse methodologies and a variety of par-
behavior in relationships, including deci- ticipant populations (for example, marital
Investment model scale 361

and nonmarital relationships, heterosexual to depart from their direct self-interest for
and gay or lesbian relationships, abusive re- the good of a relationship is termed trans-
lationships; e.g., Bui, Peplau, & Hill, 1996; formation of motivation (Holmes, 1981;
Buunk, 1987; Cox, Wexler, Rusbult, & Kelley & Thibaut, 1978). Commitment ap-
Gaines, 1997; Drigotas & Rusbult, 1992; pears to play a key role in inducing benevo-
Duffy & Rusbult, 1986; Felmlee, Sprecher, lent, pro-relationship transformation. In-
& Bassin, 1990; Gelles, 1980;Hill, Rubin, & deed, strong commitment has been shown
Peplau, 1976; Kurdek, 1991, 1993; Lin & to promote a variety of relationship main-
Rusbult, 1995; Lund, 1985; Rusbult, 1980a, tenance behaviors, including (a) tendencies
1983; Rusbult, Johnson, & Morrow, 1986; to accommodate rather than retaliate when
Rusbult & Martz, 1995; Rusbult, Verette, a partner behaves badly (Rusbult, Bisson-
Whitney, Slovik, & Lipkus, 1991; Sabatelli nette, Arriaga, & Cox, 1998; Rusbult et al.,
& Cecil-Pigo, 1985; Secord, 1983; Simpson, 1991; Wieselquist, Rusbult, Foster, & Ag-
1987;South & Lloyd, 1995;Straus & Gelles, new, 1998), (b) willingness to sacrifice oth-
1986;Strube, 1988; Strube & Barbour, 1983; erwise desirable activities when partners
Van Lange, Agnew, Harinck, & Steemers, preferences are noncorrespondent (Van
1997; Van Lange, Rusbult, Drigotas, Ar- Lange et al., 1997), (c) inclinations to dero-
riaga, Witcher, & Cox, 1997; White, 1980). gate tempting alternative partners (John-
Such findings have also been observed in son & Rusbult, 1989; Simpson, Gangestad,
nonromantic contexts-for example, in re- & Lerma, 1990), and (d) tendencies toward
search on commitment and persistence in relationship-enhancing illusion, or inclina-
friendships, in formal and informal groups, tions to perceive ones relationship as both
and in organizational settings (e.g., Farrell better than and not as bad as other relation-
& Rusbult, 1981;Kanter, 1968;Leik & Leik, ships (Murray & Holmes, 1993; Rusbult,
1976;Lin & Rusbult, 1995;Meyer & Allen, Van Lange, Yovetich, Wildschut, & Verette,
1984; Mowday, Porter, & Steers, 1982;Rus- 1998;Van Lange & Rusbult, 1995).
bult, 1980b; Rusbult & Farrell, 1983; Rus-
bult, Lowery, Hubbard, Maravankin, &
Overview of the Present Research
Neises, 1988;Staw, 1981).
Of course, persistence is a rather mini- Thus, existing research not only supports
mal requirement for relationship mainte- the claim that commitment is strengthened
nance. Partners inevitably confront situ- under conditions of high satisfaction, poor
ations that are potentially harmful to the alternatives, and sizable investments, but
longevity of their involvement-situations also demonstrates that commitment di-
in which they must solve mutual problems rectly mediates tendencies to persist in
of interdependence involving destructive relationships and to enact the sorts of main-
interaction sequences, noncorrespondent tenance behaviors outlined above. Accord-
preferences, or the existence of tempting ingly, it would seem that commitment is a
alternatives. Solving such interdependence relatively powerful motive in ongoing rela-
dilemmas typically entails some cost in the tionships. Unfortunately, no published
form of effort expenditure or departure scales exist to measure the four key con-
from ones immediate self-interest. With re- structs of the Investment Model. The pre-
peated exposure to particular classes of in- sent research attempts to remedy this state
terdependence dilemma, stable response of affairs by proffering an instrument for
orientations tend to evolve. Some individu- measuring commitment and the three bases
als routinely act in accord with their direct of dependence identified by the Investment
self-interest and behave in ways that harm Model.
their relationships, whereas other individu- The three studies described below pre-
als exhibit willingness to enact costly or ef- sent the Investment Model Scale, which in-
fortful pro-relationship behaviors. cludes measures of commitment level, satis-
The process by which individuals come faction level, quality of alternatives, and
362 C.E. Rusbult, J. M . Martz, and C.R. Agnew

investment size. We began with a version of together. For example, although depend-
the Investment Model Scale that was simi- ence and commitment grow over time, the
lar to the scales employed in previous re- mere passage of time is not sufficient to
search on the Investment Model (e.g., Rus- cause increasing commitment (i.e., some re-
bult, 1983;Rusbult et al., 1991;Van Lange et lationships develop slowly whereas others
al., 1997); seeking to develop an increas- develop quickly). Moreover, assuming that
ingly refined instrument, as we proceeded the Investment Model variables reflect dif-
through the three studies a few scale items ferences between relationships rather than
were deleted, added, or modified. In addi- differences between individuals, these vari-
tion to filling out the Investment Model ables should exhibit negligible associations
Scale, participants in Studies 2 and 3 also with personal dispositions such as self-es-
completed instruments measuring diverse teem or need for cognition.
qualities of relationships and diverse per-
sonal dispositions. In Study 3 we also ob-
Method
tained information regarding the later
status of relationships (i.e., whether the re-
Overview of the studies
lationship had persisted or ended). The
goals of the studies were (a) to evaluate the In Study 1 we administered scale items to a
internal reliability of our measures; (b) to sample of individuals who were involved in
obtain evidence regarding the convergent ongoing romantic relationships, employing
and discriminant validity of these measures; items that have been utilized in previous
and (c) to assess the predictive validity of research on the Investment Model. In
the measures. Study 2 we modified a few scale items
To examine scale reliability and validity, based on the results of Study 1, and admin-
item analyses, factor analyses, and correla- istered 12 additional instruments in order
tional analyses were performed on the data to explore the convergent and discriminant
obtained in Studies 1,2, and 3. To examine validity of Investment Model Scales. In
the convergent and discriminant validity of Study 3 we made a few final refinements of
measures, in Studies 2 and 3 we examined scale items, administered four of the valid-
the associations of Investment Model vari- ity-relevant instruments that were utilized
ables with extant instruments measuring in Study 2, and conducted follow-up tele-
several features of relationships as well as phone interviews to determine whether
several personal dispositions. Given that each relationship persisted over time and
commitment and the bases of dependence exhibited good adjustment.
emerge over the course of involvement
with a partner, these variables presumably
Participants
tell us a good deal about the nature of a
given relationship, but presumably have Study 1. Participants in Study 1 were 415
much less to do with the personal disposi- undergraduates (243 women, 172 men) who
tions of the involved persons (e.g., their per- took part in the study in partial fulfillment
sonalities). Also, given that the Investment of the requirements for introductory psy-
Model variables support persistence and chology courses at the University of North
other pro-relationship behaviors, these Carolina at Chapel Hill. Sign-up sheets
variables should exhibit moderate associa- listed a requirement for participation: To
tions with other variables reflecting supe- participate, you must be involved in a dat-
rior couple functioning, such as dyadic ad- ing relationship of at least one week in du-
justment, trust, and love. However, we ration; volunteers who were not involved
anticipated that the Investment Model vari- in dating relationships were allowed to par-
ables would be only weakly related to ticipate in an alternative project. Partici-
purely temporal features of relationships pants were 19.36 years old on average. Most
such as duration or amount of time spent were freshmen or sophomores (39% fresh-
Investment model scale 363

men, 37% sophomores, 16Y0 juniors, 8 % scribed their relationships as monogamous


seniors), and the majority were Caucasian (82% said neither partner dated others, 5%
(10% African American, 2% Asian Ameri- said one partner dated others, 13% said
can, 84% Caucasian, five percent other). both partners dated others).
Participants had been involved with their
partners for an average of 19.69 months
Procedure
( M d n = 13.50).
Studies 1 and 2. One to seven participants
Study2. A total of 326 individuals took attended each research session. The experi-
part in Study 2 in partial fulfillment of the menter described the project as a study of
requirements for introductory psychology attitudes and behavior in romantic relation-
courses. Sign-up sheets listed the same re- ships, and explained that each participant
quirement for participation as was em- would be asked to complete a computer-as-
ployed in Study 1. Thirteen individuals sisted questionnaire describing his or her
were deleted from the sample because they current romantic relationship. The ques-
had missing data for one or more variables, tionnaire was presented via personal com-
leaving 313 participants (164 women, 149 puters, linked through a server via network
men). Participants were 19.55 years old on software. The experimenter explained how
average. Most were freshmen or sopho- to use the computer, and participants pro-
mores (33% freshmen, 39% sophomores, ceeded through the questionnaire at their
18% juniors, 10% seniors), and the majority own pace. At the end of the session, partici-
were Caucasian (8% African American, pants were thoroughly debriefed and
3% Asian American, 84% Caucasian, 5% thanked for their assistance.
other). Participants had been involved with
their partners for an average of 19.09 Study3. Five to 20 participants attended
months ( M d n = 13.00). each research session. Participants (a) com-
pleted paper-and-pencil questionnaires
Study3. Participants in Study 3 were 186 measuring each Investment Model con-
individuals (96 women, 90 men) who took struct; (b) filled out forms listing their
part in the study in partial fulfillment of the names, addresses, telephone numbers, and
requirements for introductory psychology partners names or initials; and (c) indi-
courses. Sign-up sheets listed the same re- cated whether they were willing to take
quirement as was employed in Studies 1 part in a follow-up telephone interview
and 2. n o men were deleted from the sam- during the following semester; 159 partici-
ple because they described friendships pants agreed to be telephoned (85Y0 of the
rather than dating relationships. We were 186 Time 1 participants). At the end of the
able to contact 337 of the Time 1 partici- Time 1 session, participants were partially
pants (83 women, 54 men) for Time 2 fol- debriefed and thanked for their assistance.
low-up interviews. At Time 1, participants Two to 5 months after the Time 1 sessions
were 19.23 years old on average. Most were we contacted participants for follow-up in-
freshmen or sophomores (34% freshmen, terviews ( M = 15.47 weeks), attempting to
44% sophomores, 17% juniors, 5 % seniors), telephone each individual on as many as 10
and the majority were Caucasian (loo/, Af- occasions. A total of 137 participants com-
rican American, 1% Asian American, 89% pleted Time 2 interviews (86% of the 159
Caucasian, 1 % other). At Time 1, partici- Time 1 participants who agreed to be con-
pants had been involved with their partners tacted)-four individuals had moved from
for an average of 15.96 months ( M d n = the community, and we were unable to con-
13.00). Most described their relationships tact an additional 18 others. The relation-
as steady dating relationships (10% dating ships of 36 participants had ended by Time
casually, 14% dating regularly, 71% dating 2 (21 women, 15 men), and 101 of the 137
steadily, 5% engaged or married), and de- Time 2 participants were still involved with
364 C.E. Rusbult, J.M. Martz, and C.R. Agnew

their partners (62 women, 39 men). At the life. Again, note that the facet items are
end of the Time 2 interviews we mailed all included to enhance the comprehensibility
participants complete debriefing informa- of the global items, and that the global
tion. items are employed in formal tests of In-
vestment Model hypotheses. Given that the
facet items are included in the instrument
Questionnaires
solely for the purpose of improving the
Studies 1,2, and 3: Facet measures of Invest- quality of our global measures, most of the
ment Model constructs. Following a proce- analyses reported below do not include the
dure employed in previous studies (e.g., facet items.
Rusbult, 1980a, 1980b, 1983), we included
two types of items to measure satisfaction, Studies I , 2, and 3: Global measures of In-
alternatives, and investments: (a) facet vestment Model constructs. As noted ear-
items, which measure concrete exemplars lier, the questionnaire items employed in
of each construct, and (b) global items, or Study 1 were similar to those employed in
general measures of each construct (for the previous research on the Investment Model;
final version of the Investment Model these items were developed based on the
Scale, see Appendix). This approach is theoretical meaning of each Investment
based on the assumption that some partici- Model construct (e.g., Rusbult, 1980a, 1983;
pants might find it difficult to respond to Rusbult et al., 1991). In Studies 2 and 3 we
broad global items such as I have invested deleted,modified,or added a few items in an
a great deal in my relationship. Facet items effort to develop increasingly refined and
prepare participants to answer global items reliable measures of each construct. In all
by activating thoughts about each construct three studies, participants reported degree
and concretely illustrating each construct. of agreement with each item using 9-point
Thus,facet items are utilized to enhance the Likert scales (in Studies 1 and 2 , l = agree
comprehensibility of global items, thereby not at all; 9 = agree completely; in Study 3,O
increasing their reliability and validity-the = do not agree at all; 4 = agree somewhat; 8
facet items are included solely to obtain = agree completely). In Study 1,five items
good global measures of each Investment each were included for Satisfaction Level,
Model construct. The global measures of Quality of Alternatives, and Investment
each construct are the measures that are Size; in light of the centrality of the commit-
employed in formal tests of Investment ment construct in the Investment Model,
Model hypotheses. Commitment Level was measured by 12
The facet items were developed based items. Study 2 included six items for Satisfac-
on (a) previous research regarding the In- tion, six items for Alternatives, nine items
vestment Model (e.g., Rusbult, 1983; Rus- for Investments, and 11 items for Commit-
bult et al., 1991) and (b) pretesting reported ment. Study 3 included five items each for
by Drigotas and Rusbult (1992). The Satis- Satisfaction, Alternatives, and Investments,
faction Level facet items assessed the de- and nine items for Commitment.Table 1lists
gree to which the relationship gratified the the items that were selected for retention
individuals specific needs for intimacy, based on preliminary item analyses per-
companionship, sexuality, security, and formed for each study; the Appendix
emotional involvement. The Quality of Al- presents the final version of the Investment
ternatives facet items assessed the degree Model Scale. (Studies 1 and 2 included
to which each of the above needs could be scales to measure additional constructs that
fulfilled in alternative relationships (e.g., by are not relevant to the present research; ac-
another dating partner, friends, family).The cordingly, these scales will not be discussed.)
Investment Size facet items tapped in-
vested time, shared identity, shared memo- Study 2: Validity-relevantmeasures. In Study
ries, self-disclosure, and shared intellectual 2, participants completed 12 additional in-
Investment model scale 365

struments. Six of these instruments meas- the future is uncertain, I know my partner
ured qualities of ongoing relationships. will always be ready and willing to offer
These instruments were selected because me strength and support; -3 = agree not
they are relatively prominent in the close at all; +3 = agree completely; alpha = .89),
relationships literature, and because they and includes subscales to measure Predict-
sample diverse theoretical orientations ability, Dependability, and Faith (alphas =
(e.g., Self-Expansion Theory, Equity The- .74, .78, .82). The Liking and Loving Scale
ory). The 32-item Dyadic Adjustment Scale includes 18 nine-point Likert items (Rubin,
is comprised of Likert, dichotomous, and 1970; e.g., I feel that I can confide in
checklist items (Spanier, 1976; e.g., DO about virtually anything; 0 =
you kiss your partner?; 1 = never; 5 = dont agree at all; 8 = agree completely;
every day), and includes subscales to meas- alphas = .90, .88). The Equity in Relation-
ure Dyadic Consensus, Affective Expres- ship Scale includes four 8-point Likert
sion, Dyadic Satisfaction, and Dyadic Co- items (Walster, Walster, & Traupmann,
hesion (alphas = 36, .69, .85, .71). Given 1978; e.g., All things considered, how
that this scale includes four items that are would you describe your outcomes from
relevant to Satisfaction or Commitment your relationship?; -4 = extremely nega-
(e.g., How often do you discuss or have tive; +4 = extremely positive; alpha =
you considered ending your relation- .83).
ship?), we examined both (a) Total Dy- Six of the Study 2 instruments assessed
adic Adjustment, based on the original 32- personal dispositions. These instruments
item scale (alpha = .91), and (b) were selected because they sample diverse
Satisfaction- and Commitment-Purged individual-level attributes (e.g., cognitive
Adjustment, based on a 28-item scale ex- style, perceived control, self-esteem). The
cluding items that are related to either Sat- Balanced Inventory of Desirable Respond-
isfaction or Commitment (alpha = 39). ing is a 40-item instrument measuring both
The Relationship Closeness Inventory Self-Deception and Impression Manage-
measures three components of closeness; ment (Paulhus, 1991; e.g., I never cover up
this 75-item instrument includes Likert, my mistakes [reverse-scored], 1 = not
checklist, and fill-in-the-blank items (Ber- true; 7 = very true; alphas = .68, -67). The
scheid, Snyder, & Omoto, 1989; e.g., scale Multivariate Need for Cognition in-
6
does not influence my pre- cludes 25 true/false items designed to as-
sent financial security [reverse-scored]; 1 sess Cognitive Persistence, Cognitive Com-
= strongly disagree; 7 = strongly agree; al- plexity, and Cognitive Confidence (Tanaka,
pha = .44), and includes subscales to meas- Panter, & Winborne, 1988; e.g., I only
ure Frequency of Contact, Strength of In- think as hard as I have to [reverse-scored];
fluence, and Diversity of Contact (alphas 0 = false; 1 = true; K-R 20 for the total
for the former = .59, .91; K-R 20 for the scale and for subscales = .83, .77, .58, .61).
latter = .88). The scale Inclusion of Other The instrument titled Multivariate Evalu-
in the Self presents seven Venn diagrams ation of Self includes five 9-point Likert
representing varying degrees of overlap items (Hoyle, 1991;e.g., I sometimes think
between circles labeled to represent the I am a worthless individual [reverse-
self and the partner; the respondent selects scored]; 1 = not at all like me; 9 = very
the diagram that best describes the re- much like me; alpha = .93). The Affiliation
lationship (Aron, Aron, & Smollan, 1992; and Independence Inventory includes 20
choices range from completely separate, four-point Likert items (Eidelson, 1980;
nonoverlapping circles [11 to nearly com- e.g., I do not go out of my way to meet
plete overlap [7]). The Trust Scale assesses people; 1 = very uncharacteristic; 4 =
relationship-specific trust with 17 six-point very characteristic; alphas = .84, .61). The
Likert items (Rempel, Holmes, & Zanna, Collective Self-Esteem Scale taps Member-
1985; e.g., Though times may change and ship Collective Self-Esteem, Private Col-
Table 1. Means, standard deviations, item-total correlations, and alphas f o r items designed to measure each investment model construct: Studies
I, 2, and 3
Study 1 ( n = 415) Study 2 ( n = 313) Study 3 ( n = 186)
AlphdItem Alpha/Item AlphaAtem
M SD Totalr M SD Total r M SD Total r
~~~~~~~~~~~ ~~~-

Commitment Level-Global items .91 .91 .95


I am committed to maintaining my relationship with my partner. 7.59 1.88 .87 7.63 1.95 .84 6.83 1.75 .92
I want our relationship to last for a very long time. 7.53 2.07 .85 7.46 2.20 .83 6.75 1.89 .90
I feel very attached to our relationship-very strongly linked 7.42 1.94 .78 7.42 2.14 .84 6.54 1.95 .82
to my partner.
It is likely that I will date someone other than my partner 6.23 2.81 .72 6.23 2.84 .72 5.09 2.86 .75
within the next year. (-)
I would not feel very upset if our relationship were to end in 7.69 2.17 .66 6.85 1.88 .87
the near future.a (-)
g I want our relationship to last forever. 5.68 2.65 .89
Q\ I am oriented toward the long-term future of my relationship . . . 5.76 2.59 .87
Our relationship is likely to end in the near future. (-) 7.14 2.27 .73
Satisfaction Level-Global items .92 .95 .94
I feel satisfied with our relationship. 7.19 1.92 .80 7.04 2.10 .85 6.68 1.41 .88
My relationship is much better than others relationships. 7.30 1.93 .85 7.23 2.11 .87 6.51 1.72 .83
My relationship is close to ideal. 6.46 2.30 .84 6.52 2.40 .86 6.03 1.85 .86
Our relationship makes me very happy. 7.63 1.75 .87 7.38 1.99 .91 6.74 1.57 .85
Our relationship does a good job of fulfilling my needs 7.62 1.85 34 6.97 1.36 .82
for intimacy. . .
I love my partner. 7.82 2.09 .66
Quality of Alternatives-Global items .82 .85 .88
My alternatives are attractive to me (dating another, 4.71 2.58 .80 4.77 2.43 .75 4.17 2.50 .64
spending time. . .).
My alternatives to our relationship are close to ideal . . . 4.02 2.51 .70 4.01 2.39 .65 3.95 2.42 .82
If I werent dating my partner, I would do fine-I would 5.66 2.44 .62 5.73 2.18 .59 4.52 2.41 .64
find another appealing. . .
Table 1. Continued
Study 1 ( n = 415) Study 2 ( n = 313) Study 3 (n = 186)
Alphalltern AlphdItem Alpha/Item
M SD Totalr M SD Total r M SD Total r
~

The people other than my partner . . . are very appealing. 4.62 2.56 .64 5.19 2.43 55 3.94 2.59 .70
My needs for intimacy, companionship, etc., could easily 3.77 2.39 -72 3.14 2.40 .75
be fulfilled . . .
Zf I werent involved in a dating relationship, Z would 6.79 2.06 .31
enjoy spending time . . .
Investment Size-Global items .84 .84 .82
I have put a great deal into our relationship that I 6.69 2.46 .79 6.30 2.32 .75 6.08 2.07 .66
would lose.. .
Compared to other people I know, I have invested a great 6.51 2.18 .74 6.03 2.14 .70
w
Q\
deal. . .
4 I feel very involved in our relationship-like I have put a 7.25 2.06 .71 6.42 1.94 .66
great deal into it.
Many aspects of my life have become linked to my partner. .. 5.00 2.30 .65 3.89 2.44 .64
My relationships with friends and family members would 4.09 2.36 .44 3.02 2.60 .47
be complicated. . .
I have invested a great deal in our relationship that Z 6.48 2.54 .80
would lose, . .
Z have put things into our relationship that would be lost if 6.93 2.38 .59
our relationship . . .
There are special activities that are associated with 5.51 2.85 .62
our relationship . . .
There are things that are now tied to our relationship that I 3.48 2.77 .48
would lose . . .

Note: Items listed above are those that were retained based on preliminary examination of the data; items from Study 1 that were later deleted from the instrument are pre-
sented in italics. (-) Designates items that are reverse-scored. In the column labeled Alpha I Item-Total r, alpha coefficients are displayed in italics; item-total correlation coef-
ficients are not displayed in italics.
Study 3 this item was worded in the affirmative (i.e., not was deleted).
368 C.E. Rusbult, J. M . Martz, and C.R. Agnew

lective Self-Esteem, Public Collective Self- sible, (b) your partner was mainly responsi-
Esteem, and Importance to Identity using ble, or (c) you were equally responsible for
16 seven-point Likert items (Luhtanen & ending the relationship? If the participant
Crocker, 1992;e.g., In general, Im glad to reported that he or she was still dating the
be a member of the social groups I belong partner, the interviewer asked five follow-
to; -3 = strongly disagree; 3 = strongly up questions that paralleled items from the
agree; alphas for the total scale and for Dyadic Adjustment Scale (e.g., How often
subscales = .86, .75, .70, .75, .60). The Inter- do you think things are going well between
nality, Powerful Others, and Chance Scale you and your partner? Do you think things
includes 24 six-point Likert items (Leven- are going well . . .,; 1 = all the time; 2 =
son, 1981; e.g., To a great extent my life is sometimes; 3 = rarely [reverse-scored]; al-
controlled by accidental happenings; - 3 pha = .69). At the end of the interview, the
= strongly disagree; 3 = strongly agree; al- interviewer answered any questions and
phas = .62, .74, .78). thanked the participant for his or her assis-
tance.
Study 3: Validity-relevant measures. The
Study 3 questionnaire also included instru-
Results
ments to measure Dyadic Adjustment (al-
phas for the two total scores and for
Reliability analyses
subscales = .92, .90, .85, .70, .87, .74), Need
for Cognition (K-R 20 for the total scale To evaluate the reliability of our scales and
and for subscales = .84, .74, .67, .59), and to determine whether it would be desirable
both Self-Deception and Impression Man- to delete, modify, or add any items, we per-
agement (alphas = .62, .81). We included a formed reliability analyses on the items in-
25-item version of the Self-Esteem instru- cluded to measure each construct in Studies
ment, measuring not only Global Self-Es- 1,2, and 3. In a few instances we included a
teem (as in Study 2), but also Social, Physi- greater number of items than we intended
cal, Task, and Public Self-Esteem (alphas = to retain in the scale (e.g., in Study 2 we
.87, .85, .87, -80, .88). included nine items to measure Investment
Size); in these cases we performed prelimi-
nary analyses to delete poor items. The re-
Study 3: Time 2 follow-up interviews
sults of analyses for retained items are sum-
In Study 3, one of 14 trained undergraduate marized in Table 1.The alpha for each set of
research assistants contacted each partici- items is displayed in italics (see columns
pant to obtain follow-up information re- labeled Alpha/Item-Total r, statistics in
garding his or her relationship. The inter- italics); item-total correlations are also dis-
viewer reminded the individual of the study played (see Alpha/Item-Total r, nonitali-
in which he or she had previously partici- cized statistics). Any items that were in-
pated,indicating that at the time of the study cluded in Study 1 but later were deleted
the participant was dating a person named from the instrument are presented in italics.
(the interviewer read the name For each construct, the items listed for
or initials of the partner from the Time 1 Study 3 are those we recommend employ-
questionnaire). The interviewer asked ing in future research utilizing this instru-
whether the participant was still dating this ment (i.e., all nonitalicized items; see also
person (yes or no). If the participant an- the Appendix).
swered no, the interviewer said, Im sorry These analyses revealed good reliability
[with elaboration, if appropriate]. Let me for the global items designed to measure
ask just one follow-up question. Who would each construct. Alphas ranged from .91 to
you say was responsible for ending the rela- .95 for Commitment Level, .92 to .95 for
tionship? Who most wanted it to end? Satisfaction Level, .82 to -88 for Quality of
Would you say (a) you were mainly respon- Alternatives, and .82 to .84 for Investment
Investment model scale 369

Size.' We also calculated alphas for the tor in the first column, the Satisfaction fac-
facet items that were included to measure tor in the second column, and so on.
Satisfaction, Alternatives, and Investments Examination of the factor loadings for
(these analyses are not presented in Table items measuring Satisfaction, Alternatives,
1).Because the facet items tap concrete ex- and Investments revealed that for each vari-
emplars of each construct, it would not be able (a) all items loaded on a single factor
surprising if these items exhibited relatively with coefficients exceeding .40 (see coeffi-
lower reliabilities than the global items cients displayed in italics); and (b) no items
(e.g., a participant might feel satisfied with exhibited cross-factor loadings exceeding
companionship features of a relationship, an absolute value of .4O.The scale items used
yet feel dissatisfied with the level of secu- in Study 1 were refined a bit for use in Stud-
rity it provides). Nevertheless, these analy- ies 2 and 3 by deleting, adding, or modifying
ses revealed acceptable reliability for the items so as to develop an increasingly suit-
facet items included for each construct. Al- able instrument (e.g., we developed Invest-
phas ranged from .79 to -93 for the Satisfac- ment items that better represented the vari-
tion facet items, .88 to .93 for the Alterna- ety of possible investments in relationships).
tives facet items, and .73 to .84 for the In Study 1, results for the Commitment
Investments facet items. items were not as orderly. Three of the five
Commitment items loaded on a single, in-
Factor analyses dependent factor with coefficients exceed-
ing .40, but all five Commitment items ex-
We performed factor analyses to determine hibited sizable cross-loadings on the
whether the items designed to measure Satisfaction factor. However, as a conse-
each construct (a) exhibited high factor quence of a few deletions and additions to
loadings on a single factor, and (b) did not the overall scale, the analyses for Studies 2
exhibit high factor loadings for factors tap- and 3 revealed factor structures for Com-
ping other constructs. The four Investment mitment that were as orderly as were those
Model variables were expected to be corre- for Satisfaction, Alternatives, and Invest-
lated, so the analyses employed oblique, ments. To ensure that the items measuring
promax rotations. In all three studies, factor Commitment and Satisfaction were empiri-
analyses revealed four factors with eigen- cally distinguishable, we performed sepa-
values exceeding 1.OO, collectively account- rate factor analyses of the items tapping
ing for 98% to 100% of the variance in these two constructs. Once again, the analy-
scale items. A summary of the results of ses employed oblique, promax rotations.
these analyses is displayed in Table 2, which Across all three studies, the analyses re-
presents the factors in a fixed order across vealed two factors with eigenvalues exceed-
the three studies. with the Commitment fac- ing 1.00, collectively accounting for 100%
of the variance in scale items. In Study 1,all
five Commitment items loaded on a single
1. The reliability coefficients for Alternatives and In- factor, and four of the five Satisfaction
vestments were somewhat lower than were those items loaded on a single factor; one Satis-
or Commitment and Satisfaction. We have ob- faction item exhibited a cross-loading on
served this tendency in previous research regarding the Commitment factor. In Studies 2 and 3,
the Investment Model (cf. Rusbult et al., 1998; Van
Lange et al., 1997) and have speculated that the for both Commitment and Satisfaction (a)
lower reliabilities observed for these constructs all items loaded on a single factor, and (b)
may be due to the multifacetednature of Alterna- no items exhibited cross-factor loadings.
tives and Investments. For example, the Alterna- Thus, our factor-analytic findings pro-
tives items tap such diverse qualities as the prob- vide good evidence regarding the inde-
ability of finding another appealing dating partner,
the desirability of the general field of eligibles,and pendence of items designed to measure
the acceptability of spending time with friends or each Investment Model construct. In addi-
on one's own. tion. the inter-factor correlations reveal
Table 2. Factor analysis of Global Investment Model scale items: Studies 1,2, and 3
Study 1 ( n = 415) Study 2 (n = 313) Study 3 (n = 186)

F4 F1 F3 F2 F4 F1 F2 F3 F1 F2 F3 F4
Standardized Coefficients-Rotated Factor Pattern
Commitment Level-Global items
I am committed to maintaining my relationship with my partner. .49 .65 -.05 -.06 .91 .15 .02 -.11 .94 .ll .03 -.07
I want our relationship to last for a very long time. .46 .67 -.01 -.03 .76 .15 -.11 -.05 .96 .06 .05 -.02
I feel very attached to our relationship-very strongly linked. . . .41 .66 -.01 -.04 .57 .21 -.05 .22 .44 .29 -.07 .23
w It is likely that I will date someone other than my partner . . . (-) .10 SO -.32 .01 .43 .28 -.16 .02 .43 .15 -.32 .06
4
o I would not feel very upset if our relationship were to end . . .(-) .44 .10 -.14 .13 .79 .21 .06 . .03
I want our relationship to last forever. .66 .01 -.18 .17
I am oriented toward the long-term future of my relationship . .. .64 -.01 -.22 .13
Our relationship is likely to end in the near future. (-1 .09 .62 -.17 -.01

Satisfaction Level-Global items


I feel satisfied with our relationship. .03 .89 .03 -.06 .15 .87 .05 -.12 -.02 .92 -.09 b.06
My relationship is much better than others relationships. -.19 .94 .07 .08 -.04 .85 .04 .16 -.05 .86 .01 .06
My relationship is close to ideal. -.18 .92 -.02 .08 -.05 .86 -.01 .10 -.04 .83 .06 .20
Our relationship makes me very happy. .08 .91 .05 .OO .16 .82 -.02 .01 .30 .79 .02 - .19
Our relationship does a good job of fulfilling my needs . . . .01 .86 -.01 .OO .13 .75 -.03 -.01
I love my partner. .21 .58 .03 .20
Quality of Alternatives-Global items
My alternatives are attractive to me (dating another, spending , . .). .05 -.08 .89 -.03 -.04 .04 .84 -.01 -.02 -.03 .91 .08
My alternatives to our relationship are close to ideal . . . .03 -.14 .73 .OO -.01 .Q2 .73 -.02 .02 .04 .75 .04
If I werent dating my partner, I would do fine-I would find . . . -.07 .04 .63 -.02 .05 .02 .67 -.05 .12 .03 .69 -.17
The people other than my partner . . . are very appealing. -.03 -.15 .65 .03 -.23 .07 .55 .10 -.23 -.02 .62 .03
My needs for intimacy, companionship, etc., could easily be . . . .03 -.18 .71 -.06 -.06 -.lo .69 -.06
If I werent involved in a dating relationship, I would enjoy . . . .03 .ll .45 .07

Investment Size-Global items


I have put a great deal into our relationship that I would lose . . . .05 .OO -.03 .94 .02 .04 -.01 .75 .ll .03 -.15 .56
Compared to other people I know, I have invested a great deal . . . .16 -.13 .07 .88 .12 .04 .04 .78
I feel very involved in our relationship-like I have put a great . . . .22 .12 .oo .67 .19 .16 .07 .65
Many aspects of my life have become linked to my partner.. . -.I2 .09 -.08 .63 -.03 .03 -.15 .54
My relationships with friends and family members would be . . . -.14 .06 -.02 .49 -.07 -.08 .02 .57
Z have invested a great deal in our relationship that I would lose . . . .05 -.07 -.04 .97
w
I have put things into our relationship that would be lost . . . .OO .10 .13 .62
2 There are special activities that are associated with our relationship . -.08 .02 .01 .61
There are things that are now tied to our relationship that I would . . -.12 .05 -.02 .46
Eigenvalues
Eigenvalues = 1.78 34.21 4.13 13.77 2.24 38.95 5.24 3.61 65.27 6.74 4.35 3.26
Inter-Factor Correlations
Factor 1-Commitment Level .33 -.44 .38 .65 -.56 .57 .69 -.58 .63
Factor 2-Satisfaction Level -.49 .34 -.42 .53 - .42 .53
Factor 3-Quality of Alternatives - .21 - .43 -.48
Factor &Investment Size

Note: Items listed above are those that were retained based on preliminary examination of the data; items from Study 1 that were later deleted from the instrument are pre-
sented in italics. (-) Designates items that are reverse-scored. F1 = Factor 1 from the analysis, F2 = Factor 2, F3 = Factor 3, and F4 = Factor 4.All coefficients exceeding an ab-
solute value of .40 are displayed in italics.
372 C.E. Rusbult, J.M. Martz, and C.R. Agnew

that the four factors exhibit the predicted ous regression analyses to examine the
pattern of association with one another (see unique contribution of each basis of de-
bottom of Table 2). The Commitment factor pendence in predicting commitment. For
was positively correlated with the Satisfac- Study 3, we performed these analyses for
tion factor and the Investments factor, and both the full Time 1 sample and for the
was negatively correlated with the Alterna- subset of Time 1relationships that persisted
tives factor. Also, the three bases of de- through Time 2. All four analyses revealed
pendence relate to one another as would be that the three factors collectively predicted
expected-the Satisfaction factor was Commitment Level (R2s ranged from .69 to
negatively correlated with the Alternatives .77; all ps < .01). Examination of the coeffi-
factor and was positively correlated with cients for each of the three predictor vari-
the Investments factor, and the Alterna- ables revealed that 11of 12 regression coef-
tives factor was negatively correlated with ficients were significant. Satisfaction Level
the Investments factor. was positively predictive of Commitment (4
of 4 effects were significant; betas ranged
from -47 to .69),and Quality of Alternatives
Associations among measures
was negatively predictive of Commitment
The results of the above analyses revealed (4 of 4 effects were significant;betas ranged
evidence of generally good reliability and from -.29 to -.32). The coefficient for In-
validity for the set of items designed to vestment Size was nonsignificant in Study 1,
measure each Investment Model construct. but in the remaining instances, Investment
Accordingly, a single measure of each con- Size was positively predictive of Commit-
struct was formed by averaging the items ment (3 of 4 effects were significant; the
associated with each variable. Separately significant betas ranged from .19 to .27).
for Studies 1,2, and 3, correlational analy- These analyses provide good support for
ses were performed to ensure that these the Investment Model hypothesis that Sat-
variables exhibited the anticipated pattern isfaction, Alternatives, and Investments ac-
of associations. Results of these analyses count for independent variance in commit-
are displayed in Table 3. ment.
Three features of these results are note-
worthy. First, the analyses revealed evi-
Correlations with other features of
dence of acceptable convergent and dis-
relationships and with personal dispositions
criminant validity: Correlations of the facet
measures with the global measures re- To explore broader issues of validity, we
vealed that each facet measure was more performed correlational analyses to exam-
powerfully correlated with its correspond- ine the associations of Investment Model
ing global measure than with global meas- variables with measures of six additional
ures of other constructs (see values in ital- features of relationships, with the duration
ics)-this was true for the facet and global of the relationship, and with measures of six
measures of Satisfaction (rs for Studies 1,2, personal dispositions (along with subscales
and 3 = .87, .90, .83), Alternatives (rs = -73, for each instrument). Results of these
.76, .62), and Investments (TS = .67,.85, .78). analyses are displayed in Table 4.
Second, and consistent with Investment
Model hypotheses, Commitment Level was Correlations with other features of ongoing
significantly positively correlated with Sat- relationships. First, we reviewed correla-
isfaction, negatively correlated with Alter- tions of Investment Model variables with
natives, and positively correlated with In- other features of ongoing relationships. As-
vestments. And third, the analyses revealed suming that the Investment Model vari-
only moderate collinearity among the three ables support persistence and other pro-re-
Investment Model bases of dependence. lationship behaviors, we anticipated that
We performed three-factor simultane- these variables would exhibit moderate as-
Table 3. Correlations among all Global and Facet Investment Model Scales: Studies 1,2, and 3
Study 1 ( n = 415) Study 2 ( n = 313) Study 3 (n = 186)

COM SAT ALT INV COM SAT ALT INV COM SAT ALT INV

Commitment Level-Global (COM) .84** -.62** .33** .75** -.60** .60** .75** -.66** .68**
Satisfaction Level-Global (SAT) -.46** .35** -.42** .55** -.45** .54**
w Quality of Alternatives-Global (ALT) -.17** -.41** -.49**
4
W Investment Size-Global (INV)
Satisfaction Level-Facet Items ,63** .87** -.40** .33** .67** .90** -.38** .SO** .a** .83** -.42** .58**
Quality of Alternatives-Facet Items -.61** -.53** .73** -.26** -.62** -.54** .76** -.45** -.57** -.43** .62** -.42**
Investment Size-Facet Items .56** .61** -.32** .67** .68** .62** -.46** .85** .73** .65** -.48** .78**

*p< .05. **p<.O1.


Table 4. Correlations of Global Investment Model Scales with measures of other features of ongoing relationships and with measures of
personal dispositions: Studies 2 and 3

Study 2 ( n = 313) Study 3 (n = 186)


COM SAT ALT INV COM SAT ALT INV
~ ~ ~

Correlations with Other Features of Ongoing Relationships


Dyadic Adjustment
Total Adjustment Score .56** .68** -.31** .24** .69** .71** -.46** .32**
Satisfaction- and Commitment-Purged Score so** .60** -.27** .21** .63** .66** -.41** .28**
Dyadic Consensus .37** .36** -.23** .18** .45** .49** -.25** .15*
Affective Expression .38** so** -.23** .16** .40** .44** -.31** .10
Dyadic Satisfaction .63* * .79** -.41** .33** .76** .78** -.51** .45**
Dyadic Cohesion .38** .51** -.22** .27** .48** so** -.26** .31**
w Time 2 Dyadic Adjustment .40** .52** -.34** .22*
rl
P Relationship Closeness
Total Closeness Score .30** .29** -.25** .36**
Frequency of Contact .08 .10 - .12* .14*
Diversity of Contact .19** .18** -.16** .20**
Strength of Influence .51** .46** -.33** .54**
Inclusion of Other in the Self .67** .69** - .46** .55**
Trust Level
Total Trust Score .43* * .61** -.19** .28**
Predictability .35** .47** -.20** .07
Dependability .39** .58** -.19** .18**
Faith .57** .71** -.35** .34**
Liking and Loving
Liking for Partner .51** .64** -.26** .29**
Love for Partner .75** .71** -.46** .65**
Equity in Relationship -.07 .15* -.02 - .04
Duration of Relationship .15** .09 - .05 .26** .23** .04 -.lo .36**
Correlations with Personal Dispositions
Socially Desirable Responding
Self-Deception .02 .13* .17* .05 .15* .25** .04 .06
Impression Management .17* .15* -.17* .14* .24** .20** - .09 .12
Need for Cognition
Cognitive Persistence .oo -.01 .oo .03 .02 .04 .oo -.07
Cognitive Complexity -.lo -.13* .04 - .09 - .06 .03 .08 -.11
Cognitive Confidence .06 .08 .04 - .01 .08 .14* .07 - .05
Self-Esteem
Global Self-Esteem .ll .15* .06 -.03 .06 .15* .10 - .05
Social Self-Esteem -.07 .09 .27** -.12
Physical Self-Esteem .oo .07 .08 .oo
Task Self-Esteem .10 .21** .04 - .03
Public Self-Esteem - .10 .01 .21** -.15*
Affiliation and Independence Needs
Need for Affiliation .oo .08 .ll -.02
Need for Independence - .06 -.16* .03 .02
Collective Self-Esteem
Total Collective Self-Esteem .12 .19** .02 .09
Membership Collective Self-Esteem .09 .17* .06 .08
Private Collective Self-Esteem .07 .07 .oo .06
Public Collective Self-Esteem .06 .16* .03 .04
Importance to Identity .13* .12* .01 .05
Internality, Powerful Others, and Chance Control
Internality .12* .17** .oo .01
Powerful Others Control - .08 - .06 .08 .02
Chance Control -.14* -.17** .03 -.11*

*p<.05. **p<.Ol.
376 C. E. Rusbult, J.M. Martz, and C.R. Agnew

sociations with other variables reflecting in the Self measure (4 of 4 effects were
superior couple functioning-that is, with significant), for Trust Level and the Trust
variables listed under Correlations with subscales (15 of 16 effects were significant),
Other Features of Ongoing Relationships and for measures of Liking and Love for
such as Dyadic Adjustment and Inclusion the Partner (8 of 8 effects). However, the
of Other in the Self (Table 4). At the same measure of Equity in Relationship was sig-
time, we anticipated that the Investment nificantly correlated with only one Invest-
Model variables would be only weakly re- ment Model variable, Satisfaction Level;
lated to purely temporal features of rela- that is, Investment Model variables are
tionships such as duration or amount of largely independent of the degree to which
time spent together. partners perceive that they receive out-
In Study 2, we calculated correlations of comes from their relationship that are com-
the four Investment Model variables with mensurate with their inputs.
six relationship-level instruments and the Moreover, although Total Relationship
subscales associated with those instru- Closeness scores were significantly associ-
ments-as well as with Duration of Rela- ated with the Investment Model variables
tionship-for a total of 76 separate analy- (4 of 4 effects), these findings were ac-
ses. In Study 3, we calculated correlations of counted for largely by links with the
the four Investment Model variables with Strength of Influence subscale (4 of 4 ef-
the questionnaire measure of Dyadic Ad- fects). The Investment Model variables
justment and the subscales of that instru- were largely independent of the relatively
ment, with the measure of Dyadic Adjust- concrete aspects of closeness tapped by this
ment obtained in Time 2 telephone instrument-the Investment Model vari-
interviews, and with Duration of Relation- ables were weakly correlated with the Di-
ship, for a total of 32 separate analyses. As versity of Contact subscale, and were essen-
anticipated, the Investment Model vari- tially unrelated to the Frequency of
ables exhibited moderate to strong associa- Contact subscale. Finally, although Dura-
tions with most indices of superior couple tion of Relationship was weakly positively
functioning. Of the 108 correlational analy- correlated with Commitment Level and In-
ses examining the associations of Invest- vestment Size (4 of 4 effects), Duration was
ment Model variables with other features not significantly associated with either Sat-
isfaction Level or Quality of Alternatives.
of ongoing relationships, 97 effects were
statistically significant.
Stronger commitment and greater de- Correlations with personal dispositions. We
pendence on a relationship-that is, higher now consider correlations of the Invest-
Satisfaction, poorer Alternatives, and ment Model variables with measures of di-
greater Investments--consistently were as- verse personal dispositions. Assuming that
sociated with superior functioning in rela- the Investment Model variables reflect dif-
tionships. The Investment Model variables ferences between relationships rather than
consistently were associated with general differences between individuals, we antici-
Dyadic Adjustment, with the Adjustment pated that these variables would exhibit
subscales, and with the follow-up measure negligible associations with personal dispo-
of Adjustment obtained in Study 3 Time 2 sitions-that is, with the variables listed in
interviews: Stronger Commitment, greater Table 4 under Correlations with Personal
Satisfaction, poorer Alternatives, and Dispositions. In Study 2, we calculated
greater Investments were linked with correlations of the four Investment Model
higher levels of Dyadic Adjustment (51 of variables with six dispositional measures
52 effects were significant). A similar pat- and the subscales associated with those in-
tern was evident for the Inclusion of Other struments, for a total of 64 separate analy-
Investment model scale 377

ses. In Study 3, we calculated correlations of women vs. men) on the Investment Model
the Investment Model variables with three variables and measures of Dyadic Adjust-
dispositional measures and the subscales ment obtained in Study 3. Results of these
associated with those instruments, for a to- analyses are summarized in Table 5 (see
tal of 40 separate analyses. Consistent with columns labeled Ended and Persisted,
expectations, the Investment Model vari- along with accompanying Fs).
ables exhibited relatively weak associations Consistent with expectations, in com-
with personal dispositions. Of the 104 parison to relationships that later ended,
analyses performed on the data obtained in individuals in relationships that persisted
Studies 2 and 3, a total of 29 effects were reported significantly stronger Time 1
statistically significant. Importantly, only 2 Commitment, greater Satisfaction, poorer
of 104 correlations exceeded an absolute Alternatives, and greater Investment Size.
value of .25. Also, in comparison to relationships that
The Investment Model variables were ended, relationships that persisted exhib-
essentially unrelated to Need for Cognition ited significantly greater Time 1Dyadic Ad-
(only 2 of 24 effects were significant) and justment. But although the relationships
Affiliation and Independence Needs (only that persisted versus ended differed, the in-
1 of 8 effects), and were weakly associated dividuals involved in those relationships
with Self-Esteem (6 of 24 effects were sig- did not: Consistent with expectations, out of
nificant), Collective Self-Esteem (5 of 20 10 analyses examining personal disposi-
effects), and Internality, Powerful Others, tions, only one effect was significant (ac-
and Chance Control (5 of 12 effects). The cordingly, results for personal dispositions
Investment Model variables exhibited sev- are not displayed in Table 5)-compared to
eral significant correlations with Self-De- individuals in relationships that ended, indi-
ception and Impression Management (10 of viduals in relationships that persisted ex-
16 effects), but one effect was opposite to hibited significantly lower Time 1Cognitive
that which would be expected, and only one
Complexity.
of these effects exceeded an absolute value
Time 1 Commitment Level was particu-
of .25. Thus, commitment and the three
larly robust in predicting the later status of
bases of dependence-as measured by the
a relationship. We performed three simulta-
Investment Model Scale-are generally in-
neous regression analyses, regressing
dependent of a wide range of personal dis-
breakup status onto (a) Commitment Level
positions.
alone, (b) a three-factor model including
Satisfaction, Alternatives, and Investments,
and (c) a four-factor model including Com-
Predicting breakup status mitment, Satisfaction, Alternatives, and In-
Predicting Time 2 persisted versus ended vestments. Model comparison tests (Cra-
status. In Study 3, we obtained an addi- mer, 1972) revealed that the four-factor
tional measure that is crucial to evaluating model was not significantly superior to the
the validity of our instrument: In Time 2 one-factor model including only Commit-
follow-up telephone interviews we deter- ment Level ( R ~ = s .27 vs. .23; F[3, 1321 =
mined whether each participants relation- 2.41, ns). Indeed, Commitment alone pre-
ship persisted or ended during the time that dicted breakup status as powerfully as did
elapsed since Time 1 participation. To de- the three-factor model including the three
termine whether earlier self-reports of bases of dependence (R2s = .23 vs. .21). In
Commitment, Satisfaction, Alternatives, the four-factor regression model, the coeffi-
and Investments predict the later status of cient for Commitment was significant (beta
a relationship, we performed two-way = .46) but the coefficients for Satisfaction
analyses of variance (persisted vs. ended, and Investments were not (betas = .ll and
Table 5. Means and inferential statistics for the Investment Model Scales and the Dyadic Adjustment Scale as a function of persisted versus
ended status: Study 3

Ended Persisted Leavers Abandoned Stayers


n = 36 n = 101 F n = 30 n=6 n = 101 F

Associations with Investment Model Scales


Commitment Level 4.65 6.79 41.76** 4.36 6.97 6.79 24.87**
Satisfaction Level 5.56 6.96 29.61** 5.61 5.86 6.96 12.07**
2 Quality of Alternatives 4.56 3.84 5.56" 4.67 3.40 3.84 2.48'
Investment Size 4.02 5.43 19.90** 4.00 4.78 5.43 9.71**
Associations with Dyadic Adjustment
Total Adjustment Score 137.33 149.87 16.34** 136.90 140.90 149.87 7.51""
Satisfaction- and Commitment-Purged Score 118.99 128.88 14.41** 118.78 120.40 128.88 7.17**
Dyadic Consensus 58.91 63.09 8.16** 58.86 58.90 63.09 4.19*
Affective Expression 12.57 13.75 6.61** 12.45 13.40 13.75 3.88*
Dyadic Satisfaction 45.24 50.75 21.49** 45.21 47.40 50.75 9.40**
Dyadic Cohesion 20.38 22.07 7.93** 20.10 21.20 22.07 3.81*

Note: For analyses contrasting persisted versus ended relationships, df = 1,133:for analyses contrasting stayers, abandoned, and voluntary leaven, df = 2,131.
+p<.lO. *p<.05. **p<.Ol.
Investment model scale 379

.O9); the coefficient for Alternatives was Only six individuals described themselves
significant but positive (beta = .19; this ef- as abandoned, so analyses based on this
fect was opposite in direction to that which distinction are somewhat tenuous in the
would be predicted, presumably due to a present research. Nevertheless, two-way
suppressor effect). Thus, and consistent analyses of variance (stayer vs. abandoned
with Investment Model predictions, Com- vs. leaver, women vs. men) revealed find-
mitment is a powerful predictor of persisted ings that paralleled those reported above.
versus ended status, plausibly mediating the Results of these analyses are summarized in
effects on breakup of the three bases of Table 5 (see columns labeled Leavers,
dependence (Baron & Kenny, 1986). Abandoned, and Stayers, along with
We also performed two additional two- accompanying Fs).
factor simultaneous regression analyses, re- Voluntary leavers, the abandoned, and
gressing breakup status onto (a) Commit- voluntary stayers differed with respect to
ment and Total Dyadic Adjustment, as well Time 1 Commitment, Satisfaction, Alterna-
as onto (b) Commitment and the Satisfac- tives, and Investments (the effect for Alter-
tion- and Commitment-Purged measure of natives was marginal). Once again, Com-
Adjustment. In these analyses, the coeffi- mitment was particularly robust in
cients for Commitment were significant differentiating between voluntary leavers
(betas = S O and .49), but the coefficients and voluntary stayers. Compared to volun-
for the respective measures of Adjustment tary leavers and voluntary stayers, the
were not (betas = .OO and .02). In addition, abandoned exhibited a pattern that might
we performed tests of the significance of be termed entrapment-they experi-
difference between dependent effects to di- enced low satisfaction, but had invested at
rectly compare the strength of the Commit- a moderate level and possessed very poor
ment-breakup association to the strength of alternatives, and accordingly exhibited
association between breakup and each of commitment that was as strong or stronger
the measures of Dyadic Adjustment (Co- than that of voluntary stayers. These find-
hen & Cohen, 1983). These tests revealed ings replicate previous results regarding
that the association of Commitment with the relationship between breakup respon-
breakup was significantly stronger than any sibility and Investment Model variables
of the associations of Dyadic Adjustment (Rusbult, 1983). As was observed for per-
with breakup (for the strongest challenger, sisted versus ended status, leavers, the
Dyadic Satisfaction, t = 2.14, p<.Ol).Thus, abandoned, and stayers differed signifi-
Commitment Level appears to be a more cantly with respect to Time 1 Dyadic Ad-
powerful predictor of breakup status than justment (all six effects were significant;
dyadic adjustment. see Table 5).2 But importantly, although the
relationships of leavers, the abandoned, and
Predicting responsibility for breakup. In ad- stayers differed, the individuals involved in
dition to ascertaining persisted versus those relationships did not differ: Out of 10
ended status, for relationships that ended analyses examining personal dispositions,
we assessed responsibility for the
breakup-whether the breakup was volun-
tary (leaver; the participant was mainly 2. We do not report follow-up regression analyses,
mediation analyses, and effect size comparisons for
responsible or it was mutual) or nonvolun- the stayer versus abandoned versus leaver distinc-
tary (abandoned; the partner was mainly tion because (a) the sample size was small for the
responsible). Investment Model variables abandoned category, and (b) the results largely
should predict breakup status more effec- parallel those reported above for persisted versus
tively for voluntary leavers than for ended status (e.g., in comparing the effect size for
Commitment to that of the strongest Dyadic Ad-
nonvoluntary leavers, or for individuals justment predictor, the predictive power of Com-
who were abandoned by their partners mitment was significantly 5tronger; t = 2.86,
(Drigotas & Rusbult, 1992; Rusbult, 1983). pc.01).
380 C.E. Rusbult, J.M. Martz, and C.R. Agnew

only one effect was significant (accordingly, women tend to exhibit stronger commit-
results for the personal dispositions are not ment and greater dependence (higher satis-
displayed in Table 5)-leavers, the aban- faction, poorer alternatives, greater invest-
doned, and stayers differed in Cognitive ments) than do men.4
Complexity.3
Discussion
Sex differences in model variables
Reliability and validity of the Investment
We also examined mean differences be-
Model Scale
tween women and men in their responses to
Investment Model Scale items. In Studies 1 Three studies provided evidence regarding
and 2, we performed one-way analyses of the reliability and validity of the Invest-
variance (ANOVAs, women vs. men); in ment Model Scale, an instrument devel-
Study 3, we examined sex effects in the con- oped to measure commitment level and
text of two-way ANOVAs (persisted vs. three bases of dependence-satisfaction
ended, women vs. men). These analyses re- level, quality of alternatives, and invest-
vealed that, in comparison to men, women ment size. To begin with, the instrument ap-
exhibited higher Satisfaction in Study 3 pears to have good internal structure: The
(Study 3 M s = 3.46 vs. 3.64; F [l, 1331 = items designed to measure each construct
4.48, p<.05); lower Quality of Alternatives exhibit good reliability,with high item-total
in Studies 1 and 2 (Study 1 Ms = 5.41 vs. correlations and strong alpha coefficients.
4.98; F [l, 4121 = 5.32,~<.05;Study 2 M s = In general, factor analyses revealed four
5.01 vs. 4.40; F [l, 3021 = 8.42, ~K.01); factors with no substantial cross-factor
greater Investments in Study 3 (Study 3 M s loadings. Thus, the Investment Model
= 2.95 vs. 3.29; F [l,1331 = 8,66,p=.Ol);and subscales have good internal consistency,
stronger Commitment in Studies 1 and 2 and the instrument appears to measure four
(Study 1 M s = 6.95 vs. 7.35; F [l, 4121 = independent constructs.
4.51,~<.05;Study 2 M s = 7.03 VS. 7.52; F[1, In addition, we obtained good evidence
3101 = 4.86,~<.05).In Studies 1 and 2,men regarding the convergent and discriminant
also reported greater Facet Alternatives validity of the scale. Given that commit-
than did women (Study 1M s = 4.96 vs. 4.40; ment and the three bases of dependence
F [l, 4121 = 6.35,~<.01;Study 2 M S = 4.81 are assumed to develop over the course of
vs. 3.98; F [l, 3011 = 10.55, p<.Ol). These involvement with a specific partner-and
differences are relatively small in an abso- given that these variables have been shown
lute sense, and are consistent with previous to support persistence and other pro-rela-
findings (e.g., Duffy & Rusbult, 1986;Lin & tionship behaviors-we expected that the
Rusbult, 1995; Rusbult et al., 1998): Pre- Investment Model variables would exhibit
vious studies have demonstrated that sex moderate associations with other variables
differences in the Investment Model vari- reflecting superior couple functioning. As
ables are somewhat unreliably observed, anticipated, commitment and the three
but that when women and men d o differ, bases of dependence were moderately asso-
ciated with dyadic adjustment (Spanier,
3. The interaction of participant sex with persisted
versus ended status was significant for the facet 4. In Study 2, in comparison to men, women reported
measures of Alternatives and Investments: In rela- greater Predictability and Faith on the Trust Scale,
tionships that persisted, women exhibited poorer along with greater Liking for their partners. In
facet Alternatives than did men; in relationships Study 3, women scored higher on the Affective
that ended, women and men exhibited equivalent Expression subscale of Dyadic Adjustment than
facet Alternatives. In relationships that persisted, did men. Women also exhibited less Self-Decep-
women and men exhibited equivalent facet Invest- tion, greater Need for Affiliation, greater Impor-
ments; in relationships that ended, women re- tance of Collective Identity, and greater Private,
ported greater facet Investments than did men. Public, and Total Collective Self-Esteem.
Investment model scale 381

1976), with the strength of influence tionships that persisted over time, relation-
subscale of the Relationship Closeness In- ships that ended exhibited the predicted
ventory (Berscheid, Snyder, & Omoto, pattern of Time 1 scores-relationships that
1989),and with inclusion of other in the self ended by Time 2 exhibited lower Time 1
(Aron et al., 1992), trust level (Rempel et satisfaction, superior Time 1 alternatives,
al., 1985), and both liking and love for the and lower Time 1 investment size. Impor-
partner (Rubin, 1970). tantly, individuals in relationships that
At the same time, we anticipated that ended by Time 2 reported substantially
responses to the instrument would not be weaker Time 1commitment. Indeed, model
unduly colored by individual differences. comparison analyses and effect size com-
As noted earlier, the explanatory basis of parisons revealed that, in predicting later
the interdependence orientation rests on an relationship status, commitment not only
analysis of the interdependence structure outperformed the three investment model
characterizing a given relationship, not on variables, but also exceeded the predictive
the personal dispositions of the involved power of the Dyadic Adjustment Scale
persons. Accordingly, we expected that the (Spanier, 1976), a frequently-employed cri-
Investment Model variables would exhibit terion for assessing quality of couple func-
relatively weak associations with instru- tioning. Parallel results were obtained when
ments measuring a variety of personal dis- we examined responsibility for breakup, in
positions. Consistent with expectations, analyses comparing voluntary stayers, the
commitment and the three bases of de- abandoned (individuals whose partners
pendence were unrelated to---or only ended the relationship), and voluntary leav-
weakly related to-need for cognition ers (individuals who themselves ended the
(Tanaka et al., 1988), affiliation and inde- relationship).
pendence needs (Eidelson, 1980), self-es- We also examined sex differences in lev-
teem (Hoyle, 1991), and perceived control els of Investment Model variables. Sex dif-
over outcomes (Levenson, 1981). In addi- ferences were inconsistently observed, but
tion, the four subscales were only weakly we obtained some evidence that, in com-
associated with measures of self-deception parison to men, women reported higher sat-
and impression management (Paulhus, isfaction, poorer alternatives, greater in-
1991); these associations were weak in an vestments, and stronger commitment. As
absolute sense, and this is not particularly noted earlier, some previous tests of the
surprising in light of the demonstrated asso- investment model have revealed sex differ-
ciations of commitment and the bases of ences; most such studies have not. The stud-
dependence with tendencies toward posi- ies obtaining evidence of sex differences
tive illusion (e.g., Rusbult et al., 1998). consistently have revealed findings such as
Study 3 also produced data regarding ac- those uncovered in the present research
tual persistence in relationships, in order to (e.g., Duffy & Rusbult, 1986;Lin & Rusbult,
assess the predictive validity of the Invest- 1995; Rusbult et al., 1998). Thus, to the ex-
ment Model Scale. Two to 5 months follow- tent that women and men differ in mean
ing measurement of each Investment levels of model variables, women are likely
Model construct, we conducted telephone to exhibit greater dependence and stronger
interviews to determine whether each rela- commitment than are men.
tionship persisted over time, and to obtain It is appropriate to comment on three
measures of later adjustment for those rela- specific findings from analyses examining
tionships that endured. Among relation- the convergent and discriminant validity of
ships that persisted, Time 2 adjustment was our subscales: First, it is interesting that
positively correlated with Time 1 commit- commitment and the bases of dependence
ment, satisfaction, and investments, and was by and large were unrelated to a measure of
negatively correlated with Time 1quality of equity in the relationship (Walster et al.,
alternatives. Also, in comparison to rela- 1978). Satisfaction was weakly associated
382 C. E. Rusbult, J.M. Martz, and C.R. Agnew

with equity level, but commitment, alterna- tions for definitions of closeness based on
tives, and investments were virtually unre- such concrete, behaviorally based features
lated to the equity in a relationship. These of involvement (cf. Berscheid et al., 1989;
findings suggest that considerations of fair- Kelley et al., 1983). We believe that such
ness-or at least, considerations of equity definitions constitute an overly literal rep-
per se-may be largely irrelevant to the de- resentation of interdependence constructs.
velopment of dependence and commitment At the very least, in conceptualizing close-
in an ongoing relationship. It might be fruit- ness it seems important to take into account
ful to explore such issues in future research, such qualities as the degree to which shared
especially in light of the fact that the inter- activities occur in life domains that are im-
dependence and equity orientations fre- portant to the involved partners, or the ex-
quently are (incorrectly) perceived to ad- tent to which long-duration involvements
vance parallel hypotheses. effectively gratify partners needs (cf. Drig-
Second, although the strength of influ- otas & Rusbult, 1992). More generally, we
ence subscale of the Relationship Close- suspect that, rather than devoting our en-
ness Inventory (Berscheid et al., 1989) ergy to examining the concrete features of
exhibited a moderate association with com- interaction, it may be more fruitful to focus
mitment and the bases of dependence, on the stable motives that reliably emerge
these variables were only weakly associated out of circumstances of interdependence,
with the diversity of contact subscale, and including phenomena such as commitment
were unrelated to the frequency of contact and trust (cf. Holmes & Rempel, 1989;Rus-
subscale. It appears that although commit- bult et al., 1994; Rusbult, Wieselquist, Fos-
ted partners exert fairly strong influence ter, & Witcher, in press-b).
over one anothers lives, they do not neces-
sarily engage in a wide range of activities
Tests of Investment Model predictions
together, nor do they necessarily spend
enormous amounts of time together. Such Consistent with previous research examin-
findings suggest that the experience of ing the validity of Investment Model hy-
commitment and the state of dependence potheses, commitment exhibited the pre-
have a good deal to do with the broad ef- dicted associations with both global and
fects that partners exert on one anothers facet measures of three bases of depend-
lives, and have much less to do with their ence (for reviews of this literature, see Rus-
day-to-day behavioral togetherness. bult, 1987;Rusbult & Buunk, 1993; Rusbult
Third, commitment and investment size et al., 1994): Commitment was positively
were weakly positively associated with the associated with satisfaction level, nega-
duration of relationships. Thus, commit- tively associated with quality of alterna-
ment and investments unfold over time in tives, and positively associated with invest-
the expected manner-roughly speaking, ment size. Indeed, all three bases of
these variables exhibit a broad tendency to- dependence accounted for unique variance
ward cumulative accrual. At the same time, in commitment level. Thus, once again we
satisfaction level and quality of alternatives find that commitment is strengthened to
were unrelated to the duration of relation- the degree that an individual more power-
ships. Consistent with our findings for fre- fully depends on a relationship-to the ex-
quency of contact, these variables are tent that the individual wants to persist with
largely unrelated to purely temporal fea- a partner (experiences high satisfaction),
tures of relationships-the mere passage of feels bound to persist (has invested a good
time is not sufficient to cause increasing deal), and has no choice but to persist (pos-
satisfaction or declining alternatives. sesses poor alternatives).
Our findings for duration of involve- Also, analyses predicting breakup status
ment, frequency of contact, and diversity of revealed that the three bases of depend-
shared activities have important implica- ence were significantly predictive of
Investment model scale 383

breakup. However, the collective effects of global measures of dependence tend to


these three variables were no stronger than range from .55 to .85; e.g., Lin & Rusbult,
the effect on breakup of commitment level 1995; Rusbult et al., in press). Therefore, in
alone. Indeed, mediation analyses revealed administering the four-variable Investment
evidence congruent with the claim that Model Scale, we believe that it is advisable
commitment level mediates the effects on to include both facet and global measures
breakup of satisfaction, alternatives, and in- of the three bases of dependence (see Ap-
vestments. These findings are consistent pendix). The facet items should be included
with the results of previous studies demon- to enhance measurement quality for the
strating that, in predicting persistence and global items; the global items should be em-
other pro-relationship behaviors, commit- ployed in formal analyses involving Invest-
ment partially or wholly mediates the ef- ment Model variables (i.e., administer both
fects of the three bases of dependence (e.g., facet and global items; analyze only global
Rusbult, 1983; Rusbult et al., 1998; Van items).
Lange et al., 1997b). At the same time, it is entirely appropri-
Moreover, analyses examining responsi- ate to utilize the commitment subscale on
bility for breakup revealed that, in compari- its own, in the absence of either facet or
son to voluntary leavers and voluntary stay- global measures of the bases of depend-
ers, the abandoned exhibited a pattern that ence. Earlier, we described commitment as
might be termed entrapment, or nonvol- the psychological construct that influences
untary dependence (cf. Thibaut & Kelley, everyday behavior in relationships-as the
1959):The abandoned exhibited low Time 1 psychological experience that dependent
satisfaction, but had invested at a moderate individuals carry around with them. Ac-
level and possessed poor alternatives, and cordingly, it is relatively easy to access self-
accordingly reported feelings of commit- reported commitment, Indeed, in light of
ment that were as strong or stronger than the strength of association between com-
that reported by voluntary stayers. These mitment and breakup, commitment level
findings replicate previous results regard- arguably is an excellent single indicator of
ing the associations of Investment Model overall couple adjustment. (Recall that in
variables with breakup responsibility (e.g., predicting breakup, commitment level out-
Rusbult, 1983). For better or worse, strong performed the Dyadic Adjustment Scale,
commitment demonstrably promotes per- an industry standard for measuring qual-
sistence in a relationship. ity of couple functioning.)
A second user issue concerns the partici-
pant population for which the Investment
Recommendations for using the
Model Scale is suitable. The present studies
Investment Model Scale
were based on samples of North American
Before closing, we should comment on college students who were involved in on-
three important user issues. First, how going dating relationships. Clearly, the dat-
should future researchers make use of the ing relationships of North American col-
Investment Model Scale? In particular, is it lege students are not representative of all
necessary to include the facet measures of extant romantic relationships. However,
the three bases of dependence, or can the relatively good reliability and validity has
global measures stand on their own? In been obtained when versions of this instru-
some research on the Investment Model we ment were employed in research examining
have employed a version of this instrument both marital relationships and gay and les-
that excluded the facet items and observed bian relationships (e.g., Duffy & Rusbult,
reliabilities for the bases of dependence 1986; Rusbult et al., 1998), as well as in re-
that were somewhat lower than those ob- search conducted in the Netherlands and in
tained in the present work (for example, Taiwan (e.g., Lin & Rusbult, 1995; Van
without the facet items, alphas for the Lange et al., 1997b).Thus, we believe that it
384 C.E. Rusbult, J. M . Martz, and C.R. Agnew

is appropriate to employ the instrument in suitable means of measuring Investment


research involving a wide range of par- Model constructs.
ticipant populations. In doing so, it would
be desirable to tailor the instrument to the
Conclusions
population under consideration (e.g., re-
word references to dating partners in re- Three studies demonstrated the reliability
search concerning marital relations). and validity of the Investment Model Scale,
A third user issue concerns the fields providing evidence of the convergent, dis-
long-standing attachment to behaviorally criminant, and predictive validity of this in-
based measures-a seduction that presum- strument. Investment Model variables are
ably has its origins in the behavioral tradi- associated with the variables with which
tion in psychology. Should we be troubled they should be associated, and are unre-
by the fact that the Investment Model Scale lated to the variables to which they should
is a self-report instrument? The litany of be unrelated. In addition, measures of In-
our concerns with such measures is well-re- vestment Model variables obtained at ear-
hearsed, including worries that self-report lier research occasions effectively predict
measures may be colored by the desire to relationship outcomes measured at later re-
appear consistent or to present oneself fa- search occasions. Previous research using
vorably. Previous research on the Invest- similar instruments has demonstrated that
ment Model and related phenomena has commitment is a robust predictor not only
demonstrated that self-reported commit- of persistence, but also of pro-relationship
ment is associated not only with (a) the motivation and willingness to exert effort
probability that a relationship will persist, or endure cost for the good of a relation-
but also with a variety of highly specific ship (e.g., willingness to accommodate, to
behavioral indices, such as (b) the amount sacrifice, to drive away tempting alterna-
of physical effort individuals are willing to tives).
exert for a partner by stepping up and down From a theoretical point of view, it is
a stairstep (i.e., willingness to sacrifice), (c) noteworthy that the Investment Model is
the positivity of interaction behavior, as embedded in Interdependence Theory, a
coded from videotaped couple conversa- prominent orientation for understanding in-
tions (i.e., tendencies to accommodate), (d) terpersonal motivation and behavior. The
inclinations to cognitively disparage appli- Investment Model extends fundamental in-
cants for a dating service (i.e., derogation terdependence premises and constructs in
of alternatives), and (e) tendencies to em- such a manner as to illuminate our under-
ploy plural pronouns in describing ones standing of persistence and healthy func-
relationship (i.e., cognitive interdepend- tioning in ongoing relationships. The bene-
ence; e.g., Agnew et al., 1998; Drigotas & fits of its embeddedness in Interdependence
Rusbult, 1992; Felmlee et al., 1990;Johnson Theory should be clear: Researchers who
& Rusbult, 1989; Rusbult, 1983; Rusbult et seek to understand commitment processes
al., 1998;Rusbult & Martz, 1995; Rusbult et can readily link with the broader theory,
al., 1991; Simpson, 1987; Van Lange et al., benefitting from the richness and compre-
1997b). Thus, it should be clear that the In- hensiveness of an interdependence analysis.
vestment Model constructs represent more We eagerly await further research on
than purely in the head phenomena. commitment processes, encouraging close-
More generally, it is important to ask how relationships theorists and researchers to
one could measure an inherently subjective move beyond the rather exclusive focus on
construct such as commitment other than positivity of affect (i.e., satisfaction, attrac-
via self-report. Thus, although we continue tion, love) that traditionally has charac-
to believe in the desirability of behavioral terized our field. Clearly,feeling happy with
measurement to augment self-report in- a relationship is a good thing; at the same
struments, self-report methods are a highly time, happiness and positive affect are not
Investment model scale 385

by any means the whole picture when it do not. We hope that the existence of the
comes to understanding persistence and Investment Model Scale will promote fu-
pro-relationship motivation. It is by moving ture research on commitment processes,
beyond an exclusive focus on positivity of thereby extending the interdependence ori-
affect that we can begin to more fully un- entation to understanding a variety of cen-
derstand how and why some relationships tral processes in ongoing close relation-
persist and thrive over time whereas others ships.

References
Agnew, C. R., Van Lange, F? A. M., Rusbult, C. E., & commitment in homosexual and heterosexual rela-
Langston, C. A. (1997).Cognitive interdependence: tionships. Journal of Homosexuality, 12,l-23.
Commitment and the mental representation of Eidelson, R. J. (1980).Interpersonal satisfaction and
close relationships, Journal of Personality and So- level of involvement: A curvilinear relationship.
cial Psychology, 74, 939-954. Journal of Personality and Social Psychology, 39,
Aron, A., Aron, E. N., & Smollan, D. (1992).Inclusion 460-470.
of other in the self-scale and the structure of inter- Farrell, D., & Rusbult, C. E. (1981).Exchange variables
personal closeness. Journal of Personality and So- as predictors of job satisfaction,job commitment,
cial Psychology, 63,596-612. and turnover: The impact of rewards, costs, alterna-
Baron, R. M.,& Kenny, D. A. (1986).The modera- tives, and investments. Organizational Behavior
tor-mediator variable distinction in social psycho- and Human Performance, 27,7&95.
logical research Conceptual, strategic, and statisti- Felmlee, D., Sprecher, S., & Bassin, E. (1990).The dis-
cal considerations. Journal of Personality and solution of intimate relationships:A hazard model.
Social Psychology, 51,1173-1182. Social Psychology Quarterly, 53,13-30.
Becker, H. S. (1960).Notes on the concept of commit- Gelles, R. J. (1980).Violence in the family:A review of
ment. American Journal of Sociology, 66,3240. research in the seventies.Journal of Marriage and
Berscheid, E.(1994).Interpersonal relationships. An- the Family, 42,873485.
nual Review of Psychology, 45,79-129. Hill, C. T., Rubin, Z . , & Peplau, L. A. (1976).Breakups
Berscheid, E.,& Reis,H.T. (1998).Attraction and close before marriage: The end of 103 affairs. Journal of
relationships. In D. T. Gilbert, S. T. Fiske, & G. Social Issues, 32,147-168.
Lindzey (Eds) Handbook of socialpsychology (4th Holmes, J. G. (1981).The exchange process in close
ed.;Vol.2,pp. 193-281).New York:Random House. relationships:Microbehavior and macromotives. In
Berscheid, E., Snyder, M., & Omoto, A. M. (1989).The M. J. Lerner & S. C. Lerner (Eds), The justice mo-
tive in social behavior (pp. 261-284). New York:
relationship closeness inventory: Assessing the
Plenum Press.
closeness of interpersonal relationships.Journal of
Holmes, J. G., & Rempel, J. K. (1989).Trust in close
Personality and Social Psychology, 57,792-807. relationships. In C. Hendrick (Ed.), Review ofper-
Brickman, I?, Dunkel-Schetter, C., & Abbey, A. (1987).
sonality and social psychology (Vol. 10, pp.
The development of commitment. In P.Brickman
187-220).London: Sage.
(Ed.), Commitment, conflict, and caring (pp. Hoyle, R. H. (1991).Evaluating measurement models
145-221). Englewood Cliffs, NJ: Prentice-Hall. in clinical research: Covariance structure analysis of
Bui,K. T., Peplau, L. A,, & Hill, C. T. (1996).Testing the latent variable models of self-conception.Journal of
Rusbult model of relationship commitment and Consulting and Clinical Psychology,59,67-76.
stability in a 15-yearstudy of heterosexual couples. Johnson, D. J., & Rusbult, C. E. (1989).Resisting temp-
Personality and Social Psychology Bulletin, 22, tation: Devaluation of alternative partners as a
12441257. means of maintaining commitment in close rela-
Buunk, B. (1987).Conditions that promote breakups as tionships. Journal of Personality and Social Psy-
a consequence of extradyadic involvements Jour- chology, 57,967-980.
nal ofSocial and Clinical Psychology,5,271-284. Johnson, M. P (1991).Commitment to personal rela-
Cohen, J., & Cohen, P.(1983).Applied multiple regres- tionships. In W. H. Jones & D. W. Perlman (Eds),
siodcorrelation analysis for the behavioral sciences. Advances in personal relationships (Vol. 3, pp.
Hillsdale, NJ: Erlbaum. 117-143). London: Jessica Kingsley.
Cox, C. L., Wexler, M. O., Rusbult, C. E., & Gaines, Kanter, R. M. (1968).Commitment and social organi-
S. O., Jr. (1997).Prescriptive support and commit- zation: A study of commitment mechanisms in uto-
ment processes in close relationships. Social Psy- pian communities. American Sociological Review,
chology Quarterly, 60,79-90. 33,499-517.
Cramer, E. M. (1972). Significance tests and tests of Kelley,H. H. (1979).Personal relationships: Their struc-
models in multiple regression. American Statisti- tures and processes. Hillsdale, NJ: Erlbaum.
cian, 26,2630. Kelley, H. H. (1983).Love and commitment. In H. H.
Drigotas, S. M., & Rusbult, C. E. (1992).Should I stay Kelley, E. Berscheid, A. Christensen, J. H. Harvey,
.,
or should I go?: A deDendence model of breakuus. T. L. Huston, G. Levinger, E. McClintock, L. A.
Journal of Personality and Social Psychology, 62, Peplau, & D. R. Peterson (Eds.), Close relationships
62-87. (pp. 265-314).New York: W. H. Freeman.
Duffy, S., & Rusbult, C. E. (1986). Satisfaction and Kelley, H. H., Berscheid, E., Christensen, A., Harvey,
386 C.E. Rusbult, J. M. Martz, and C.R. Agnew

J. H., Huston, T. L., Levinger, G., McClintock, E., nal of Personality and Social Psychology, 16,
Peplau, L. A., & Peterson, D. R. (Eds.) (1983). 265-273.
Close relationships. New York: W. H. Freeman. Rusbult, C. E. (1980a). Commitment and satisfaction
Kelley, H. H., & Thibaut, J. W. (1978). Interpersonal in romantic associations: A test of the investment
relations: A theory of interdependence. New York: model. Journal of Experimental Social Psychology,
Wiley. 16,172-186.
Kurdek, L. A. (1991). Correlates of relationship satis- Rusbult, C. E. (1980b). Satisfaction and commitment
faction in cohabiting gay and lesbian couples: Inte- in friendships. Representative Research in Social
gration of contextual, investment, and problem- Psychology, 11,96-105.
solving models. Journal of Personality and Social Rusbult, C. E. (1983). A longitudinal test of the invest-
Psychology, 61,910-922. ment model: The development (and deterioration)
Kurdek, L. A. (1993). Predicting marital dissolution: A of satisfaction and commitment in heterosexual in-
five-year prospective longitudinal study of newly- volvements. Journal of Personality and Social Psy-
wed couples. Journal of Personality and Social Psy- chology, 45,101-1 17.
chology, 64,221-242. Rusbult, C. E. (1987). Commitment in close relation-
Leik, R. K., & Leik, S. K. (1976). Transition to interper- ships: The investment model. In L. A. Pepiau, D. 0.
sonal commitment. In R. L. Hamblin & J. H. Sears, S. E. Taylor, & J. L. Freedman (Eds.), Read-
Kenkel (Eds.), Behavioral theory in sociology (pp. ings in socialpsychology: Classic and contemporary
299-322). New Brunswick, NJ: Transaction Books. contributions (pp. 147-1.57). Englewood Cliffs, NJ:
Levenson, H. (1981). Differentiating among internal- Prentice-Hall.
ity, powerful others, and chance. In H. M. Lefcourt Rusbult, C. E.,Bissonnette, V. I., Arriaga, X. B., & Cox.
(Ed.), Research with the locus of control construct C. L. (1998). Accommodation processes during the
(Vol. 1, pp. 15-63). New York: Academic Press. early years of marriage. In T.N. Bradbury (Ed.), The
Levinger, G. (1979). A social exchange view on the developmental course of marital dysfunction (pp.
dissolution of pair relationships. In R. L. Burgess & 74-113). New York: Cambridge University Press.
T. L. Huston (Eds.), Social exchange in developing Rusbult, C. E., & Buunk, B. P. (1993). Commitment
relationships (pp. 169-193). New York: Academic processes in close relationships: An interdepend-
Press. ence analysis. Journal of Social and Personal Rela-
Lin, Y. H. W., & Rusbult, C. E. (1995). Commitment to tionships, 10,175-204.
dating relationships and cross-sex friendships in Rusbult, C. E., Drigotas,S. M.,& Verette, J. (1994).The
America and China: The impact of centrality of investment model: An interdependence analysis of
relationship, normative support, and investment commitment processes and relationship mainte-
model variables. Journal of Social and Personal Ru- nance phenomena. In D. Canary & L. Stafford
lationships, 12,7-26. (Eds.), Communication and relational maintenance
Luhtanen, R.. & Crocker, J. (1992). A collective self-es- (pp. 115-139). New York: Academic Press.
Rusbult, C. E., & Farrell, D. (1983). A longitudinal test
teem scale: Self-evaluation of ones social identity.
of the investment model: The impact on job satis-
Personulity and Social Psychology Bulletin, 18,
faction, job commitment, and turnover of vari-
302-318.
ations in rewards, costs, alternatives, and invest-
Lund, M. (1985). The development of investment and
ments. Journal of Applied Psychology, 68,429-438.
commitment scales for predicting continuity of
Rusbult, C . E., Johnson, D. J., & Morrow, G. D. (1986).
personal relationships. Journal of Social and Per- Predicting satisfaction and commitment in adult
sonal Relationships, 2,3-23. romantic involvements: An assessment of the gen-
Meyer, J. P., & Allen, N. J. (1984). Testing the side-bet eralizability of the investment model. Social Psy-
theory of organizational commitment: Some chology Quarterly, 49,8149.
methodological considerations. Journal of Applied Rusbult, C. E., Lowery, D., Hubbard, M., Maravankin,
Psychology, 69,372-378. 0. J., & Neises, M. (1988). Impact of employee
Mowday, R. T., Porter, L. W., & Steers. R. M. (1982). mobility and employee performance on the alloca-
Employee-organization linkages: The psychology tion of rewards under conditions of constraint.
of commitment, absenteeism, and turnover. New Journal o,f Personality and Social Psychology, 54,
York: Academic Press. 605-615.
Murray, S. L., & Holmes, J. G. (1993). Seeing virtues in Rusbult, C. E., & Martz, J. (1995). Remaining in an
faults: Negativity and the transformation of inter- abusive relationship: An investment model analy-
personal narratives in close relationships. Journal sis of nonvoluntary commitment. Personality and
of Personality and Social Psychology, 65,707-722. Social Psychology Bulletin, 21,558-571.
Paulhus, D. L. (1991). Measurement and control of Rusbult, C. E., Van Lange, I? A. M., Yovetich, N. A.,
response bias. In J. P. Robinson, P. R. Shaver, & Wildschut, T. & Verette, J. (1998). A functional
L. S. Wrightsman (Eds.), Measures of personality analysis of perceived superiority in close relation-
and social psychological attitudes (Vol. 1, pp. ships. Unpublished manuscript, University of
17-59). San Diego: Academic Press. North Carolina at Chapel Hill, Chapel Hill, NC.
Rempel, J. K., Holmes, J. G., & Zanna, M. P. (1985). Rusbult, C. E., Verette, J., Whitney, G. A,, Slovik. L. F.,
Trust in close relationships. Journal of Personality & Lipkus, I. (1991). Accommodation processes in
and Social Psychology, 49,95-112. close relationships: Theory and preliminary empiri-
Rubin, J. Z., & Brockner, J. (1975). Factors affecting cal evidence. Journal of Personality and Social Psy-
entrapment in waiting situations: The Rosencrantz chology, 60,53-78.
and Guildenstern effect. Journal of Personality and Rusbult, C. E., Wieselquist, J., Foster, C. A,, & Witcher,
Social Psychology, 31.1054-1063. B. S. (in press). Commitment and trust in close
Rubin, 2.(1970). Measurement of romantic love. Jour- relationships: An interdependence analysis. In
Investment model scale 387

W. H. Jones & J. M. Adams (Eds.), Conimitment in leave an abusive relationship: Economic depend-
close personal relationships: Theory and research. ence and psychological commitment. Journal of
New York: Plenum Press. Marriage and the Family, 45,785-793.
Sabatelli, R. M., & Cecil-Pigo, E. F. (1985). Relational Tanaka. J. S., Panter, A. T., & Winborne, W. C. (1988).
interdependence and commitment in marriage. Dimensions of the need for cognition: Subscales
Journal of Marriage and the Family, 47,931-937. and gender differences. Multivariate Behavioral
Secord, P. F. (1983). Imbalanced sex ratios: The social Research, 23,35-50.
consequences. Personality and Social Psychology Teger, A. 1. (1980). Too much invested to quit. New
Bulletin, 9,525-543. York: Pergamon Press.
Simpson, J. A. (1987). The dissolution of romantic rela- Thibaut, J. W., & Kelley, H. H. (1959). The social psy-
tionships: Factors involved in relationship stability chology of groups. New York: Wiley.
and emotional distress. Journal of Personality and Tropper, R. (1972). The consequences of investment in
Sociul Psychology, 53,683-692. the process of conflict. Journal of Conflict Resolu-
Simpson, J. A., Gangestad. S. W., & Lerma, M. (1990). tion, 16,97-98.
Perccption of physical attractiveness: Mechanisms Van Lange, P. A. M., Agnew, C. R., Harinck, S., &
involved in the maintenance of romantic relation- Steemers, G. (1997). From game theory to real life:
ships. Journal of Personality and Social Psychology, How social value orientation affects willingness to
59,1192-1201. sacrifice in ongoing relationships. Journal of Per-
South, S. J., & Lloyd, K. M. (1995). Spousal alternatives sonality and Social Psychology, 73,133C-1344.
and marital dissolution. American Sociological Re- Van Lange, P. A. M., & Rusbult, C. E. (1995). My
view, 60,21-35. relationship is better than-and not as bad
Spanier, G. B. (1976). Measuring dyadic adjustment: as-yours is: The perception of superiority in close
New scales for assessing the quality of marriage relationships. Personality and Social Psychology
and similar dyads. Journal of Marriage and the Bulletin, 2 1 , 3 2 4 .
Family, 38, 15-28. Van Lange, I? A. M., Rusbult, C . E., Drigotas, S. M..
Staw, B. M. (1976). Knee-deep in the big muddy: A Arriaga, X. B., Witcher, 3. S.. & Cox, C. L. (1997).
study of escalating commitment to a chosen course Willingness to sacrifice in close relationships. Jour-
of action. Organizational Behavior and Human nal of Personality and Social Psychology, 72,1373-
Performance, 1 6 , 2 7 4 . 1395.
Staw, B. M. (1981). The escalation of commitment to a Walster, E., Walster, G. W., & Traupmann, J. (1978).
course of action. Academy of Management Review, Equity and premarital sex. Journal o,f Personality
6,577-587. and Social Psychology, 36,82-92.
Straus, M. A., & Gelles, R. J. (1986). Societal change White, G. L. (1980). Physical attractiveness and court-
and change in family violence from 1975 to 1985 as ship progress. Journal of Personality and Social
revealed by two national surveys. Journal of Mar- Psychology, 39,660-668.
riage and the Family, 48,465479. Wieselquist, J.. Rusbult, C. E., Foster, C. A., & Agnew,
Strube, M. J. (1988). The decision to leave an abusive C. R. (1998). Commitment and trust in close rela-
relationship: Empirical evidence and theoretical is- tionships. Unpublished manuscript. University of
sues. Psychological Bulletin, 104,23&250. North Carolina at Chapel Hill. Chapel Hill, NC.
Strube, M. J., & Barbour, L. S. (1983). The decision to
388 C.E. Rusbult, J. M . Martz, and C.R. Agnew

Appendix

Satisfaction Level Facet and Global Items


1. Please indicate the degree to which you agree with each of the following statements regarding
your current relationship (circle an answer for each item).
(a) My partner fulfills my needs for Dont Agree Agree Agree Agree
intimacy (sharing personal thoughts, At All Slightly Moderately Completely
secrets, etc.)
(b) My partner fulfills my needs for Dont Agree Agree Agree Agree
companionship (doing things together, At All Slightly Moderately Completely
enjoying each others company, etc.)
(c) My partner fulfills my sexual needs Dont Agree Agree Agree Agree
(holding hands, kissing, etc.) At All Slightly Moderately Completely
(d) My partner fulfills my needs for Dont Agree Agree Agree Agree
security (feeling trusting, comfortable At All Slightly Moderately Completely
in a stable relationship, etc.)
(e) My partner fulfills my needs for Dont Agree Agree Agree Agree
emotional involvement (feeling At All Slightly Moderately Completely
emotionally attached, feeling good
when another feels good, etc.)
2. I feel satisfied with our relationship (please circle a number).
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely
3. My relationship is much better than others relationships.
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely
4. My relationship is close to ideal.
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely
5. Our relationship makes me very happy.
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely
6. Our relationship does a good job of fulfilling my needs for intimacy, companionship, etc.
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely

Quality of Alternatives Facet and Global Items


1. Please indicate the degree to which you agree with each statement regarding the fulfillment of
each need in alternative relationships (e.g., by another dating partner, friends, family).
(a) My needs for intimacy (sharing Dont Agree Agree Agree Agree
personal thoughts, secrets, etc.) At All Slightly Moderately Completely
could be fulfilled in alternative
relationships
Investment model scale 389

(b) My needs for companionship (doing Dont Agree Agree Agree Agree
things together, enjoying each others At AU Slightly Moderately Completely
company, etc.) could be fulfilled in
alternative relationships
(c) My sexual needs (holding hands, Dont Agree Agree Agree Agree
kissing, etc.) could be fulfilled in At All Slightly Moderately Completely
alternative relationships
(d) My needs for security (feeling Dont Agree Agree Agree Agree
trusting, comfortable in a stable At All Slightly Moderately Completely
relationship, etc.) could be fulfilled
in alternative relationships
(e) My needs for emotional involvement Dont Agree Agree Agree Agree
(feeling emotionally attached, At All Slightly Moderately Completely
feeling good when another feels
good, etc.) could be fulfilled in
alternative relationships
2. The people other than my partner with whom I might become involved are very appealing
(please circle a number).
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely
3. My alternatives to our relationship are close to ideal (dating another, spending time with friends
or on my own, etc.).
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely
4. If I werent dating my partner, I would do fine-I would find another appealing person to date.
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely
5. My alternatives are attractive to me (dating another, spending time with friends or on my own,
etc.).
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely
6. My needs for intimacy, companionship, etc., could easily be fulfilled in an alternative relationship.
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely

Investment Size Facet and Global Items


1. Please indicate the degee to which you agree with each of the following statements regarding
your current relationship (circle an answer for each item).
(a) I have invested a great deal of time in Dont Agree Agree Agree Agree
our relationship At All Slightly Moderately Completely
(b) I have told my partner many private Dont Agree Agree Agree Agree
things about myself (I disclose secrets At All Slightly Moderately Completely
to him/her)
390 C.E. Rusbult, .I.M.Martz, and C.R. Agnew

(c) My partner and I have an intellectual Dont Agree Agree Agree Agree
life together that would be difficult to At All Slightly Moderately Completely
replace
(d) My sense of personal identity Dont Agree Agree Agree Agree
(who I am) is linked to my partner At All Slightly Moderately Completely
and our relationship
(e) My partner and I share many Dont Agree Agree Agree Agree
memories At All Slightly Moderately Completely
2. I have put a great deal into our relationship that I would lose if the relationship were to end
(please circle a number).
0 1 2 3 4 5 6 7 8
D o Not Agree Agree Agree
At All Somewhat Completely
3. Many aspects of my life have become linked to my partner (recreational activities, etc.), and I
would lose all of this if we were to break up.
0 1 2 3 4 5 6 7 8
D o Not Agree Agree Agree
At All Somewhat Completely
4. I feel very involved in our relationship-like I have put a great deal into it.
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely
5. My relationships with friends and family members would be complicated if my partner and I
were to break up (e.g., partner is friends with people I care about).
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely
6. Compared to other people I know, I have invested a great deal in my relationship with my
partner.
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely

Commitment Level Items


1. I want our relationship to last for a very long time (please circle a number).
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely
2. I am committed to maintaining my relationship with my partner.
0 1 2 3 4 5 6 7 8
D o Not Agree Agree Agree
At All Somewhat Completely
3. I would not feel very upset if our relationship were to end in the near future.
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely
Investment model scale 391

4. It is likely that I will date someone other than my partner within the next year.
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely
5. I feel very attached to our relationship-very strongly linked to my partner.
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely
6. I want our relationship to last forever.
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely
7. I am oriented toward the long-term future of my relationship (for example, I imagine being with
my partner several years from now).
0 1 2 3 4 5 6 7 8
Do Not Agree Agree Agree
At All Somewhat Completely

Anda mungkin juga menyukai